[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245678#comment-17245678
]
Takanobu Asanuma commented on HDFS-15240:
-
Thanks again, [~marvelrock] and [~ferhui]!
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.2.2, 3.3.1, 3.4.0, 3.1.5
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.1.001.patch, HDFS-15240-branch-3.2.001.patch,
> HDFS-15240-branch-3.3-001.patch, HDFS-15240-branch-3.3.001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245641#comment-17245641
]
Hui Fei commented on HDFS-15240:
[~marvelrock] Thanks for your patch, [~tasanuma] Thanks for review ! commit to
branch-3.1 and branch-3.3
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.2.2, 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.1.001.patch, HDFS-15240-branch-3.2.001.patch,
> HDFS-15240-branch-3.3-001.patch, HDFS-15240-branch-3.3.001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245144#comment-17245144
]
Hui Fei commented on HDFS-15240:
Failed tests are unrelated, they passed locally. +1 on
[^HDFS-15240-branch-3.1.001.patch] [^HDFS-15240-branch-3.3.001.patch] . Will
commit tomorrow if no other comments.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.2.2, 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.1.001.patch, HDFS-15240-branch-3.2.001.patch,
> HDFS-15240-branch-3.3-001.patch, HDFS-15240-branch-3.3.001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245042#comment-17245042
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 29m
7s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} branch-3.3 Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 4m
5s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 34m
50s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
37s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 3m
24s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
53s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
27m 49s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
34s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
38s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 6m
20s{color} | {color:green}{color} | {color:green} branch-3.3 passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
25s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
0s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
37s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 15m
37s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
44s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
44s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
16m 26s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
57s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
41s{color} | {color:green}{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} || ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 10m 6s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/340/artifact/out/patch-unit-hadoop-common-project_hadoop-common.txt{color}
| {color:red} hadoop-common in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red}123m 27s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/340/artifact/out/patch-unit-hadoop-hdfs-project_hadoop-hdfs.txt{color}
| {color:red} hadoop-hdfs in the patch passed.
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244960#comment-17244960
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
26s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} branch-3.1 Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
1s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 20m
34s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 14m
41s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
30s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
31s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
19m 10s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
1s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 2m
56s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m
55s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
18s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m
55s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 14m
4s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 14m
4s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
30s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
32s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 19s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
1s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
19s{color} | {color:green}{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} || ||
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 9m
7s{color} | {color:green}{color} | {color:green} hadoop-common in the patch
passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red}109m 29s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/339/artifact/out/patch-unit-hadoop-hdfs-project_hadoop-hdfs.txt{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
52s{color} |
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244918#comment-17244918
]
Hui Fei commented on HDFS-15240:
[~tasanuma] Thanks for review! Because branch-3.3 test is not triggered, Rename
patch name, from HDFS-15240-branch-3.1-001.patch to
HDFS-15240-branch-3.1.001.patch, from HDFS-15240-branch-3.3-001.patch to
HDFS-15240-branch-3.3.001.patch, And try it again.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.2.2, 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.1.001.patch, HDFS-15240-branch-3.2.001.patch,
> HDFS-15240-branch-3.3-001.patch, HDFS-15240-branch-3.3.001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244910#comment-17244910
]
Takanobu Asanuma commented on HDFS-15240:
-
+1 on [^HDFS-15240-branch-3.1-001.patch] and [^HDFS-15240-branch-3.3-001.patch].
[~ferhui] Could you double check and commit them?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.2.2, 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.2.001.patch, HDFS-15240-branch-3.3-001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244406#comment-17244406
]
Xiaoqiao He commented on HDFS-15240:
cherry-pick to branch-3.2.2.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.2.001.patch, HDFS-15240-branch-3.3-001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244379#comment-17244379
]
HuangTao commented on HDFS-15240:
-
I have check and run the failed unit test
TestReconstructStripedFile.testNNSendsErasureCodingTasks in branch-3.1, it can
pass locally
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.2.001.patch, HDFS-15240-branch-3.3-001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244348#comment-17244348
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 19m
21s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} branch-3.1 Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 3m
30s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 27m
18s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
5s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
30s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
35s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
19m 10s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
4s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 2m
57s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m
58s{color} | {color:green}{color} | {color:green} branch-3.1 passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
18s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m
54s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 14m
4s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 14m
4s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
28s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
33s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 4s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
2s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
8s{color} | {color:green}{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} || ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 8m 42s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/332/artifact/out/patch-unit-hadoop-common-project_hadoop-common.txt{color}
| {color:red} hadoop-common in the patch failed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red}100m 35s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/332/artifact/out/patch-unit-hadoop-hdfs-project_hadoop-hdfs.txt{color}
| {color:red} hadoop-hdfs in the patch failed.
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244245#comment-17244245
]
HuangTao commented on HDFS-15240:
-
Upload [^HDFS-15240-branch-3.1-001.patch] for branch-3.1 and
[^HDFS-15240-branch-3.3-001.patch] for branch-3.3.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Blocker
> Fix For: 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.1-001.patch,
> HDFS-15240-branch-3.2.001.patch, HDFS-15240-branch-3.3-001.patch,
> HDFS-15240.001.patch, HDFS-15240.002.patch, HDFS-15240.003.patch,
> HDFS-15240.004.patch, HDFS-15240.005.patch, HDFS-15240.006.patch,
> HDFS-15240.007.patch, HDFS-15240.008.patch, HDFS-15240.009.patch,
> HDFS-15240.010.patch, HDFS-15240.011.patch, HDFS-15240.012.patch,
> HDFS-15240.013.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244195#comment-17244195
]
Takanobu Asanuma commented on HDFS-15240:
-
BTW, I'm going to change this issue to blocker. All new releases of branch-3.x
should include it.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.2.001.patch, HDFS-15240.001.patch,
> HDFS-15240.002.patch, HDFS-15240.003.patch, HDFS-15240.004.patch,
> HDFS-15240.005.patch, HDFS-15240.006.patch, HDFS-15240.007.patch,
> HDFS-15240.008.patch, HDFS-15240.009.patch, HDFS-15240.010.patch,
> HDFS-15240.011.patch, HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244187#comment-17244187
]
Takanobu Asanuma commented on HDFS-15240:
-
[~hexiaoqiao] I confirmed TestReconstructStripedFile succeeded with or without
ISA-L for branch-3.2.2 with [^HDFS-15240-branch-3.2.001.patch] in my
environment. +1 on [^HDFS-15240-branch-3.2.001.patch]. Please go ahead. Thanks.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0, 3.2.3
>
> Attachments: HDFS-15240-branch-3.2.001.patch, HDFS-15240.001.patch,
> HDFS-15240.002.patch, HDFS-15240.003.patch, HDFS-15240.004.patch,
> HDFS-15240.005.patch, HDFS-15240.006.patch, HDFS-15240.007.patch,
> HDFS-15240.008.patch, HDFS-15240.009.patch, HDFS-15240.010.patch,
> HDFS-15240.011.patch, HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244039#comment-17244039
]
Xiaoqiao He commented on HDFS-15240:
It seems failed unit test is not related with this issue. check and run
successful at local. Just backport [^HDFS-15240-branch-3.2.001.patch] to
branch-3.2.
[~marvelrock],[~tasanuma],[~ferhui] Please give another check once have
bandwidth, if no more other comments, will backport to branch-3.2.2 and prepare
to RC. Thanks all.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240-branch-3.2.001.patch, HDFS-15240.001.patch,
> HDFS-15240.002.patch, HDFS-15240.003.patch, HDFS-15240.004.patch,
> HDFS-15240.005.patch, HDFS-15240.006.patch, HDFS-15240.007.patch,
> HDFS-15240.008.patch, HDFS-15240.009.patch, HDFS-15240.010.patch,
> HDFS-15240.011.patch, HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17244016#comment-17244016
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 18m
37s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} branch-3.2 Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 3m
46s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 26m
18s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
53s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
35s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
31s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
19m 39s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
5s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
0s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs config;
considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
5s{color} | {color:green}{color} | {color:green} branch-3.2 passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
19s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
1s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
6s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 15m
6s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
34s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
34s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 34s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
2s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
20s{color} | {color:green}{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} || ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 8m 53s{color}
|
{color:red}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/328/artifact/out/patch-unit-hadoop-common-project_hadoop-common.txt{color}
| {color:red} hadoop-common in the patch failed. {color} |
| {color:green}+1{color} | {color:green} unit {color} | {color:green}104m
46s{color} | {color:green}{color} | {color:green} hadoop-hdfs in the patch
passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
52s{color} |
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17243887#comment-17243887
]
Xiaoqiao He commented on HDFS-15240:
reopen and add 3.2.2 as another target version.
upload [^HDFS-15240-branch-3.2.001.patch] from [~marvelrock] and try to trigger
Jenkins.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240-branch-3.2.001.patch, HDFS-15240.001.patch,
> HDFS-15240.002.patch, HDFS-15240.003.patch, HDFS-15240.004.patch,
> HDFS-15240.005.patch, HDFS-15240.006.patch, HDFS-15240.007.patch,
> HDFS-15240.008.patch, HDFS-15240.009.patch, HDFS-15240.010.patch,
> HDFS-15240.011.patch, HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17243662#comment-17243662
]
Takanobu Asanuma commented on HDFS-15240:
-
Thanks for committing it, [~ferhui]!
[~marvelrock] We are maintaining branch-3.1 or later, and I think they have the
same bug. Would it also be possible to provide the patch for branch-3.2 and
branch-3.1?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17243631#comment-17243631
]
Hui Fei commented on HDFS-15240:
Cherry-pick fails because of conflicts. [~marvelrock] Could you please upload
the patch for branch-3.3 ? Thanks
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17243630#comment-17243630
]
Hui Fei commented on HDFS-15240:
Push to trunk.
Thanks again [~marvelrock] for report and fix, [~umamaheswararao] [~tasanuma]
for review , [~touchida] for test, and all here
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242842#comment-17242842
]
Takanobu Asanuma commented on HDFS-15240:
-
I confirmed that the patch also worked well with ISA-L. (CI doesn't use ISA-L
since HADOOP-17224 is reverted.)
+1 on [^HDFS-15240.013.patch]. Thanks!
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242837#comment-17242837
]
Hui Fei commented on HDFS-15240:
Thanks [~marvelrock] for report and fix, [~umamaheswararao] [~tasanuma] for
review , [~touchida] for test, and all here!
Will commit tomorrow if no other comments!
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242747#comment-17242747
]
Uma Maheswara Rao G commented on HDFS-15240:
+1 latest patch looks good to me.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242648#comment-17242648
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
18s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
23s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 22m
37s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
25s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 18m
2s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
41s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
53s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
22m 41s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
56s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
0s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
18s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
27s{color} | {color:green}{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
23s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
3s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
50s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 20m
50s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
51s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Private Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 17m
51s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
41s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
50s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
16m 54s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
|
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242388#comment-17242388
]
HuangTao commented on HDFS-15240:
-
Thanks [~tasanuma] for your careful review. I address all your comments in
[^HDFS-15240.013.patch] .
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, HDFS-15240.013.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17242355#comment-17242355
]
Takanobu Asanuma commented on HDFS-15240:
-
This bug seems to have been since Hadoop-3.0.0 was released. We saw a similar
issue a long time ago, but we were unable to find its cause. Thank you very
much for finding and fixing it, [~marvelrock]. Thanks for your reviews,
[~ferhui] and [~umamaheswararao].
The main fix of [^HDFS-15240.012.patch] looks good to me. Some minor comments
for the unit tests:
* testTimeoutReadBlockInReconstruction: Please use JIRA number for the comment.
{code:java}
- // before this fix, NPE will cause reconstruction fail(test timeout)
+ // before HDFS-15240, NPE will cause reconstruction fail(test timeout)
{code}
* assertBufferPoolIsEmpty: This line could be removed?
{code:java}
- byteBuffer = null;
{code}
* emptyBufferPool: Calling {{getBuffer}} may just be enough?
{code:java}
- ByteBuffer byteBuffer = bufferPool.getBuffer(direct, 0);
- byteBuffer = null;
+ bufferPool.getBuffer(direct, 0);
{code}
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17240532#comment-17240532
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
20s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
51s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 22m
36s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
15s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 18m
0s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
42s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
51s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
22m 48s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
57s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
58s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
18s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
30s{color} | {color:green}{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
22s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
3s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
40s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 20m
40s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 18m
1s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Private Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 18m
1s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
47s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
24s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
1s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
18m 9s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
|
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17240439#comment-17240439
]
HuangTao commented on HDFS-15240:
-
Upload [^HDFS-15240.012.patch] to fix flaky test case.
The test cases can't pass with XOR-2-1 erasure coding, because these two cases
need to allow at least two blocks to be broken.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> HDFS-15240.012.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17240416#comment-17240416
]
Hui Fei commented on HDFS-15240:
[~umamaheswararao] Could you please take a look again ? Thanks!
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch, HDFS-15240.011.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17240038#comment-17240038
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
31s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
41s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 23m
8s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
20s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 18m
0s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
42s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
51s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
22m 43s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
57s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
0s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
19s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m
31s{color} | {color:green}{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
22s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
5s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
49s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 20m
49s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
56s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Private Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 17m
56s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
41s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
48s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
16m 48s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
|
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239878#comment-17239878
]
Hui Fei commented on HDFS-15240:
[~marvelrock] Thanks. It's great, [^HDFS-15240.010.patch] looks better and
clear.
{quote}
+ } catch (InterruptedException e) {
+ }
{quote}
Maybe we should log exception message, after this +1.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239657#comment-17239657
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
23s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
50s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 24m
53s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 25m
45s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
10s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
48s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
36s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
24m 17s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
5s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
8s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
51s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 6m
20s{color} | {color:green}{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
27s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
25s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 25m
40s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 25m
40s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 23m
33s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Private Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 23m
33s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 3m
10s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
43s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
19m 10s{color} | {color:green}{color} | {color:green} patch has no errors when
building and testing our client artifacts. {color} |
|
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239639#comment-17239639
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
11s{color} | {color:blue}{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
1s{color} | {color:green}{color} | {color:green} No case conflicting files
found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green}{color} | {color:green} The patch does not contain any
@author tags. {color} |
| {color:green}+1{color} | {color:green} {color} | {color:green} 0m 0s{color}
| {color:green}test4tests{color} | {color:green} The patch appears to include 2
new or modified test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
23s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 22m
25s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
32s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 18m
0s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
44s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m
49s{color} | {color:green}{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
23m 14s{color} | {color:green}{color} | {color:green} branch has no errors when
building and testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
16s{color} | {color:green}{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
11s{color} | {color:green}{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
24s{color} | {color:blue}{color} | {color:blue} Used deprecated FindBugs
config; considering switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 6m
0s{color} | {color:green}{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
28s{color} | {color:blue}{color} | {color:blue} Maven dependency ordering for
patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
11s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 24m
16s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.9.1+1-Ubuntu-0ubuntu1.18.04 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 24m
16s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
6s{color} | {color:green}{color} | {color:green} the patch passed with JDK
Private Build-1.8.0_275-8u275-b01-0ubuntu1~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 21m
6s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
2m 50s{color} |
{color:orange}https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/318/artifact/out/diff-checkstyle-root.txt{color}
| {color:orange} root: The patch generated 1 new + 9 unchanged - 0 fixed = 10
total (was 9) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
3s{color} | {color:green}{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
1s{color} | {color:green}{color} | {color:green} The patch has no whitespace
issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239546#comment-17239546
]
HuangTao commented on HDFS-15240:
-
Upload [^HDFS-15240.010.patch] to add a new function clearFuturesAndService to
remove all stale futures from readService and clear futures.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, HDFS-15240.010.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239509#comment-17239509
]
Hui Fei commented on HDFS-15240:
Trigger jenkins
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239090#comment-17239090
]
HuangTao commented on HDFS-15240:
-
Upload [^HDFS-15240.009.patch] to address
??Either we have to make sure all stale futures removed from service when
clearing the futures. Otherwise in the current code we are assuming
service#poll will give one of the submitted current future only.??
Thanks [~umamaheswararao] for that nice advice, this make code more clear. And
thanks [~ferhui] for offline discussion.
However, I'm a little worried about whether the below code slow down the
reconstruction if the future can't finish.
{code:java}
while (!futures.isEmpty()) {
try {
Future future = readService.poll(
stripedReadTimeoutInMills, TimeUnit.MILLISECONDS);
futures.remove(future);
} catch (InterruptedException e) {
}
} {code}
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> HDFS-15240.009.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17239039#comment-17239039
]
HuangTao commented on HDFS-15240:
-
The NPE fails the reconstruction and close the striped
readers(StripedBlockReconstructor Thread). In the meanwhile, the striped reader
threads are running asynchronously.
{code:java}
reconstructor.freeBuffer(reader.getReadBuffer()){code}
The above line of code just puts the buffer into BUFFER_POOL, the striped
reader still holds the buffer. So the striped reader may store dirty data in
that buffer before calling to the below line
{code:java}
reader.freeReadBuffer()
{code}
I had written a special UT `testAbnormallyCloseDoesNotWriteBufferAgain` to
indicate how the dirty buffer is generated without any fix.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17238943#comment-17238943
]
Uma Maheswara Rao G commented on HDFS-15240:
I think the latest patch mostly looks good to me.
I think [~ferhui] asked why do we need to interchange below line:
{code}
reader.closeBlockReader();
441 reconstructor.freeBuffer(reader.getReadBuffer());
442 reconstructor.freeBuffer(reader.getReadBuffer());
442 reader.freeReadBuffer();
443 reader.freeReadBuffer();
{code}
Could you please clarify?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17232775#comment-17232775
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
47s{color} | | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | | {color:green} No case conflicting files found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | | {color:green} The patch does not contain any @author tags.
{color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | | {color:green} The patch appears to include 2 new or modified
test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
26s{color} | | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 22m
28s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 23m
27s{color} | | {color:green} trunk passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 19m
11s{color} | | {color:green} trunk passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
45s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
55s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
23m 55s{color} | | {color:green} branch has no errors when building and
testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
44s{color} | | {color:green} trunk passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
42s{color} | | {color:green} trunk passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
30s{color} | | {color:blue} Used deprecated FindBugs config; considering
switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 8m
36s{color} | | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
26s{color} | | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
58s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 21m
53s{color} | | {color:green} the patch passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 21m
53s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 19m
3s{color} | | {color:green} the patch passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 19m
3s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} blanks {color} | {color:green} 0m
0s{color} | | {color:green} The patch has no blanks issues. {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
2m 45s{color} |
[/results-checkstyle-root.txt|https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/303/artifact/out/results-checkstyle-root.txt]
| {color:orange} root: The patch generated 1 new + 27 unchanged - 0 fixed = 28
total (was 27) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
57s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
26m 39s{color} | | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
47s{color} | | {color:green} the patch passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
41s{color} | | {color:green} the patch passed
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17232608#comment-17232608
]
HuangTao commented on HDFS-15240:
-
Upload [^HDFS-15240.008.patch] for fixing checkstyle.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch, HDFS-15240.008.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17232588#comment-17232588
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Logfile || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
38s{color} | | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} || ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | | {color:green} No case conflicting files found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | | {color:green} The patch does not contain any @author tags.
{color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | | {color:green} The patch appears to include 2 new or modified
test files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 2m
3s{color} | | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 20m
20s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
2s{color} | | {color:green} trunk passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
13s{color} | | {color:green} trunk passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
32s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 4m
11s{color} | | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
23m 11s{color} | | {color:green} branch has no errors when building and
testing our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
1s{color} | | {color:green} trunk passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 4m
4s{color} | | {color:green} trunk passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
15s{color} | | {color:blue} Used deprecated FindBugs config; considering
switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 8m
1s{color} | | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} || ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
27s{color} | | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
50s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 19m
23s{color} | | {color:green} the patch passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 19m
23s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
12s{color} | | {color:green} the patch passed with JDK Private
Build-1.8.0_272-8u272-b10-0ubuntu1~18.04-b10 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 17m
12s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} blanks {color} | {color:green} 0m
0s{color} | | {color:green} The patch has no blanks issues. {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
2m 31s{color} |
[/results-checkstyle-root.txt|https://ci-hadoop.apache.org/job/PreCommit-HDFS-Build/302/artifact/out/results-checkstyle-root.txt]
| {color:orange} root: The patch generated 5 new + 27 unchanged - 0 fixed = 32
total (was 27) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 4m
8s{color} | | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
24m 50s{color} | | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
3s{color} | | {color:green} the patch passed with JDK
Ubuntu-11.0.9+11-Ubuntu-0ubuntu1.18.04.1 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 4m
6s{color} | | {color:green} the patch passed
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17232477#comment-17232477
]
HuangTao commented on HDFS-15240:
-
Thanks [~ferhui] [~umamaheswararao] for your comments. I upload
HDFS-15240.007.patch for addressing your comments.
With regard to StripingChunkReadResult.OUTDATED, I think we should keep it,
which can avoid scheduling an useless block reader.
About reader.closeBlockReader(), I add UT
`testAbnormallyCloseDoesNotWriteBufferAgain`.
{code:java}
reconstructor.freeBuffer(reader.getReadBuffer());
// thread may stop here, before closeBlockReader, StripedBlockReader may write
data into the released buffer.
reader.freeReadBuffer();
reader.closeBlockReader();{code}
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, HDFS-15240.007.patch,
> image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17230489#comment-17230489
]
Uma Maheswara Rao G commented on HDFS-15240:
HI [~marvelrock] Thanks a lot for working on this issue and thanks [~ferhui]
for reviewing this issue.
This seems to be very tricky to find and debug. NPE is causing because of stale
future coming from readService#poll. This future we already cleared as part-of
previous iterations due to timeouts. Either we have to make sure all stale
futures removed from service when clearing the futures. Otherwise in the
current code we are assuming service#poll will give one of the submitted
current future only.
Anyway your fix looks good to me and it can handle the case of stale future and
avoid NPE.
Other than the above comments from [~ferhui], I have the follwoing.
I think we can avoid making StripedReconstructor public. The suggestion is:
{code:java}
@InterfaceAudience.Private
100 public abstract class StripedReconstructor {
{code}
You can create ErasureCodingTestHelper in tests with the package
*.hadoop.hdfs.server.datanode.erasurecode and add a small static util method
to access StripedReconstructor methods. This way you can avoid this src class
public.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17225810#comment-17225810
]
HuangTao commented on HDFS-15240:
-
Thanks for your suggestion, [~ferhui] I will change that later.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17225269#comment-17225269
]
Hui Fei commented on HDFS-15240:
One more comment
* There are some sleeps for waiting as following, Could we change to
GenericTestUtils.waitFor ? It can avoid flaky tests.
{quote}
+ try {
+// new block reader created by scheduleNewRead() and the
+// following block readers will wait the timeout reader
+Thread.sleep(500);
+ } catch (InterruptedException e) {
+ }
+// try to wait the block reader to finish reading
+try {
+ Thread.sleep(2000);
+} catch (InterruptedException e) {
+ // pass
+}
+ synchronized (readerIndices) {
+if (readerIndices.get(index)) {
+ break;
+}
+ }
+ try {
+Thread.sleep(200);
+ } catch (InterruptedException e) {
+// pass
+ }
{quote}
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17225229#comment-17225229
]
Hui Fei commented on HDFS-15240:
[~marvelrock] Thanks for contribution. It overall looks good. some minor
comments
* StripingChunkReadResult.OUTDATED is not used by other codes like TIMEOUT.
StripedBlockUtil.java:301 it works well when i change *if (future != null)* to
*if (future != null && futures.containsKey(future))*
* seems that there is no need to move *reader.closeBlockReader()*, and it works
well if do not move it.
{quote}
+ reader.closeBlockReader();
reconstructor.freeBuffer(reader.getReadBuffer());
reader.freeReadBuffer();
- reader.closeBlockReader();
}
{quote}
* do we need the readIndex param? It's best to do nothing when running on
product env.
{quote}
+ /**
+ * Used as a hook to inject latency when read block
+ * in erasure coding reconstruction process.
+ */
+ public void delayBlockReader(int readerIndex) { }
+
+ /**
+ * Used as a hook to inject intercept when free the block reader buffer.
+ */
+ public void interceptFreeBlockReaderBuffer(int readerIndex) { }
{quote}
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17224408#comment-17224408
]
Hui Fei commented on HDFS-15240:
[~touchida] [~marvelrock] Thanks for check, I will take a look later.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F)
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17224396#comment-17224396
]
HuangTao commented on HDFS-15240:
-
Thanks for your work, [~touchida]. I checked again and got the same result as
you.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17223584#comment-17223584
]
Toshihiko Uchida commented on HDFS-15240:
-
@Hui Fei The unit tests without the fix failed on trunk in my environment, as
expected.
Could you try [^test-HDFS-15240.006.patch], which is a patch containing only
the unit tests of [^HDFS-15240.006.patch]?
I uploaded the result and output:
[^org.apache.hadoop.hdfs.TestReconstructStripedFile.txt] and
[^org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt].
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png,
> org.apache.hadoop.hdfs.TestReconstructStripedFile-output.txt,
> org.apache.hadoop.hdfs.TestReconstructStripedFile.txt,
> test-HDFS-15240.006.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17211704#comment-17211704
]
Hui Fei commented on HDFS-15240:
[~marvelrock] Test case you added without your fix also pass successfully.
Could you please take a look? Thanks
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17209985#comment-17209985
]
Hui Fei commented on HDFS-15240:
OK! I will take a look
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17209917#comment-17209917
]
Wei-Chiu Chuang commented on HDFS-15240:
[~ferhui] could you review this patch? Looks like a great fix from a quick look.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> HDFS-15240.006.patch, image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> EDITED for readability.
> {code:java}
> decode from block[0, 2, 3, 4, 5, 7] to generate block[1', 6', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[6] and block[6'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4
> decode from block[0, 2, 3, 4, 5, 6] to generate block[1', 7', 8']
> Check the first 131072 bytes between block[1] and block[1'], the longest
> common substring length is 65536
> CHECK AGAIN: all 27262976 bytes between block[1] and block[1'], the longest
> common substring length is 27197440 # this one
> Check the first 131072 bytes between block[7] and block[7'], the longest
> common substring length is 4
> Check the first 131072 bytes between block[8] and block[8'], the longest
> common substring length is 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
>
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17194932#comment-17194932
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
9s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} dupname {color} | {color:green} 0m
0s{color} | {color:green} No case conflicting files found. {color} |
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
4s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 23m
28s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 25m
29s{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.8+10-post-Ubuntu-0ubuntu118.04.1 {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 19m
47s{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_265-8u265-b01-0ubuntu2~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
47s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
47s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
22m 19s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
11s{color} | {color:green} trunk passed with JDK
Ubuntu-11.0.8+10-post-Ubuntu-0ubuntu118.04.1 {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
35s{color} | {color:green} trunk passed with JDK Private
Build-1.8.0_265-8u265-b01-0ubuntu2~18.04-b01 {color} |
| {color:blue}0{color} | {color:blue} spotbugs {color} | {color:blue} 3m
15s{color} | {color:blue} Used deprecated FindBugs config; considering
switching to SpotBugs. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
57s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
23s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
45s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 20m
15s{color} | {color:green} the patch passed with JDK
Ubuntu-11.0.8+10-post-Ubuntu-0ubuntu118.04.1 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 20m
15s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
35s{color} | {color:green} the patch passed with JDK Private
Build-1.8.0_265-8u265-b01-0ubuntu2~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 17m
35s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
46s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
43s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
15m 27s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:red}-1{color} | {color:red} javadoc {color} | {color:red} 0m
58s{color} | {color:red}
hadoop-hdfs-project_hadoop-hdfs-jdkUbuntu-11.0.8+10-post-Ubuntu-0ubuntu118.04.1
with JDK Ubuntu-11.0.8+10-post-Ubuntu-0ubuntu118.04.1 generated 3 new + 2
unchanged - 0 fixed = 5 total (was 2) {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 3m
35s{color} | {color:green} the patch passed with JDK Private
Build-1.8.0_265-8u265-b01-0ubuntu2~18.04-b01 {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 8m
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17179461#comment-17179461
]
HuangTao commented on HDFS-15240:
-
No problem, [~ferhui] I will work on that later
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17179393#comment-17179393
]
Fei Hui commented on HDFS-15240:
[~marvelrock] Fail to apply your patch on trunk branch. Could you please rebase
your patch on trunk?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17159804#comment-17159804
]
HuangTao commented on HDFS-15240:
-
Could you please help me review that ? [~weichiu]
I think this is a very critical bug.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17158991#comment-17158991
]
Hongbing Wang commented on HDFS-15240:
--
[~marvelrock] We have the same problem. At the same time as this problem, there
are frequent FullGCs (every few seconds). Dump and MAT found that there are
lots of ecWorker objects that almost fill the entire heap.
!image-2020-07-16-15-56-38-608.png|width=722,height=591!
look forward this patch into trunk.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch,
> image-2020-07-16-15-56-38-608.png
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17083820#comment-17083820
]
maobaolong commented on HDFS-15240:
---
What a great job you did, look forward to your patch in the trunk branch.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Fix For: 3.4.0
>
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17079067#comment-17079067
]
Wu Weiwei commented on HDFS-15240:
--
We have the same problem, Our cluster of several thousand units occasionally
has problems with corrupted EC files, this patch LGTM.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Affects Versions: 3.3.1
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17077758#comment-17077758
]
guo commented on HDFS-15240:
Nice job, LGTM.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Affects Versions: 3.3.0
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17077745#comment-17077745
]
HuangTao commented on HDFS-15240:
-
Hi watchers, anyone would like to help take a review?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Affects Versions: 3.3.0
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17075854#comment-17075854
]
HuangTao commented on HDFS-15240:
-
the failed test is unrelated with this patch. [~weichiu] PTAL, Thanks
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch, HDFS-15240.004.patch, HDFS-15240.005.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17075851#comment-17075851
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
30s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
25s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 17m
44s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
9s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
37s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 4m
3s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
20m 53s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
6s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
38s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
24s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
27s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 14m
23s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 14m
23s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
24s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
38s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 19s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
30s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
40s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 9m
7s{color} | {color:green} hadoop-common in the patch passed. {color} |
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 2m
11s{color} | {color:green} hadoop-hdfs-client in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 92m 43s{color}
| {color:red} hadoop-hdfs in the patch passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
58s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}219m 52s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.server.balancer.TestBalancer |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.8 Server=19.03.8 Image:yetus/hadoop:e6455cc864d |
| JIRA Issue | HDFS-15240 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12998977/HDFS-15240.005.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 040df997653e 4.15.0-58-generic #64-Ubuntu SMP Tue Aug 6
11:12:41 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / e6455cc |
| maven | version: Apache Maven
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17075760#comment-17075760
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
28s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
22s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 18m
56s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 16m
51s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
24s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
40s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
20m 4s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
8s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
41s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
23s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
28s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 14m
25s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 14m
25s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
2m 27s{color} | {color:orange} root: The patch generated 1 new + 27 unchanged -
0 fixed = 28 total (was 27) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
37s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
14m 18s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
51s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
33s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 9m
17s{color} | {color:green} hadoop-common in the patch passed. {color} |
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 2m
10s{color} | {color:green} hadoop-hdfs-client in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 96m 54s{color}
| {color:red} hadoop-hdfs in the patch passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
52s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}227m 25s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests |
hadoop.hdfs.server.blockmanagement.TestUnderReplicatedBlocks |
| | hadoop.hdfs.server.namenode.TestAddOverReplicatedStripedBlocks |
| | hadoop.hdfs.TestReconstructStripedFile |
| | hadoop.hdfs.TestReconstructStripedFileWithRandomECPolicy |
| | hadoop.hdfs.server.datanode.TestNNHandlesBlockReportPerStorage |
| | hadoop.hdfs.server.datanode.TestDataNodeVolumeFailureReporting |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.8 Server=19.03.8 Image:yetus/hadoop:e6455cc864d |
| JIRA Issue | HDFS-15240 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12998962/HDFS-15240.004.patch |
| Optional
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17075078#comment-17075078
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
59s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
3s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 25m
31s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 17m
19s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
36s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
27s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
21m 5s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
18s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
16s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
20s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
31s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
12s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 15m
12s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
36s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
18s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
14m 23s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
44s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
14s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 8m 55s{color}
| {color:red} hadoop-common in the patch passed. {color} |
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 2m
3s{color} | {color:green} hadoop-hdfs-client in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red}104m 52s{color}
| {color:red} hadoop-hdfs in the patch passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
48s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}243m 46s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.security.TestRaceWhenRelogin |
| | hadoop.hdfs.TestReconstructStripedFileWithRandomECPolicy |
| | hadoop.hdfs.TestReconstructStripedFile |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.8 Server=19.03.8 Image:yetus/hadoop:4454c6d14b7 |
| JIRA Issue | HDFS-15240 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12998815/HDFS-15240.003.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux cefe41f3d2c1 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19
08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality |
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17075021#comment-17075021
]
HuangTao commented on HDFS-15240:
-
fix checkstyle and the unit case
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch, HDFS-15240.002.patch,
> HDFS-15240.003.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17074946#comment-17074946
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
48s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
43s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 19m
45s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
55s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
34s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
25s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
21m 7s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
19s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
15s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
19s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m
32s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 15m
9s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 15m
9s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
2m 34s{color} | {color:orange} root: The patch generated 11 new + 27 unchanged
- 0 fixed = 38 total (was 27) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
21s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
14m 22s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
45s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
15s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 8m
58s{color} | {color:green} hadoop-common in the patch passed. {color} |
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 2m
3s{color} | {color:green} hadoop-hdfs-client in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red}105m 39s{color}
| {color:red} hadoop-hdfs in the patch passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
47s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}236m 56s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.server.namenode.TestNameNodeMXBean |
| | hadoop.hdfs.TestReconstructStripedFileWithRandomECPolicy |
| | hadoop.hdfs.TestReconstructStripedFile |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.8 Server=19.03.8 Image:yetus/hadoop:4454c6d14b7 |
| JIRA Issue | HDFS-15240 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12998786/HDFS-15240.002.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 4ad954db57d2 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19
08:06:28 UTC 2019 x86_64 x86_64
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17073396#comment-17073396
]
HuangTao commented on HDFS-15240:
-
[~weichiu] I'm working on that, I will upload another patch later
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail:
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17072030#comment-17072030
]
Wei-Chiu Chuang commented on HDFS-15240:
[~marvelrock] could you double check the patch again? there are various
warnings in the precommit. Thanks.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17068170#comment-17068170
]
Hadoop QA commented on HDFS-15240:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
33s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:red}-1{color} | {color:red} test4tests {color} | {color:red} 0m
0s{color} | {color:red} The patch doesn't appear to include any new or modified
tests. Please justify why no new tests are needed for this patch. Also please
list what manual steps were performed to verify this patch. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m
13s{color} | {color:blue} Maven dependency ordering for branch {color} |
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 17m
58s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 16m
2s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 2m
43s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m
45s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
21m 6s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 7m
14s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m
42s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m
26s{color} | {color:blue} Maven dependency ordering for patch {color} |
| {color:red}-1{color} | {color:red} mvninstall {color} | {color:red} 0m
20s{color} | {color:red} hadoop-hdfs-client in the patch failed. {color} |
| {color:red}-1{color} | {color:red} compile {color} | {color:red} 1m
42s{color} | {color:red} root in the patch failed. {color} |
| {color:red}-1{color} | {color:red} javac {color} | {color:red} 1m 42s{color}
| {color:red} root in the patch failed. {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
0m 47s{color} | {color:orange} The patch fails to run checkstyle in root
{color} |
| {color:red}-1{color} | {color:red} mvnsite {color} | {color:red} 0m
21s{color} | {color:red} hadoop-hdfs-client in the patch failed. {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:red}-1{color} | {color:red} shadedclient {color} | {color:red} 3m
0s{color} | {color:red} patch has errors when building and testing our client
artifacts. {color} |
| {color:red}-1{color} | {color:red} findbugs {color} | {color:red} 0m
19s{color} | {color:red} hadoop-hdfs-client in the patch failed. {color} |
| {color:red}-1{color} | {color:red} javadoc {color} | {color:red} 0m
16s{color} | {color:red} hadoop-hdfs-client in the patch failed. {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:green}+1{color} | {color:green} unit {color} | {color:green} 8m
49s{color} | {color:green} hadoop-common in the patch passed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 0m 19s{color}
| {color:red} hadoop-hdfs-client in the patch failed. {color} |
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 91m 16s{color}
| {color:red} hadoop-hdfs in the patch passed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
35s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}188m 1s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.server.namenode.ha.TestStandbyCheckpoints |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.8 Server=19.03.8 Image:yetus/hadoop:4454c6d14b7 |
| JIRA Issue | HDFS-15240 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12997870/HDFS-15240.001.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 4d271edefa4e 4.15.0-58-generic #64-Ubuntu SMP Tue Aug 6
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17067871#comment-17067871
]
HuangTao commented on HDFS-15240:
-
I will try to add an UT
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17067682#comment-17067682
]
Fei Hui commented on HDFS-15240:
[~marvelrock] Good Catch ! Thanks for reporting and fixing.
Could you please add UT?
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k) from
> DN directly, and choose 6(b0-b5) blocks to decode other 3(b6', b7', b8')
> blocks. And find the longest common sequenece(LCS) between b6'(decoded) and
> b6(read from DN)(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet(only show 2 of 28 cases) is my check program
> output. In my case, I known the 3th block is corrupt, so need other 5 blocks
> to decode another 3 blocks, then find the 1th block's LCS substring is block
> length - 64kb.
> It means (0,1,2,4,5,6)th blocks were used to reconstruct 3th block, and the
> dirty buffer was used before read the 1th block.
> Must be noted that StripedBlockReader read from the offset 0 of the 1th block
> after used the dirty buffer.
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17067569#comment-17067569
]
Yao Guangdong commented on HDFS-15240:
--
[~weichiu] We have the same problem. PTAL.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k), and
> choose 6(b0-b5) blocks to decode other 3(b6', b7', b8') blocks. And find the
> longest common sequenece(LCS) between b6' and b6(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet is my check program output
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17067559#comment-17067559
]
guojh commented on HDFS-15240:
--
[~surendrasingh] We have the same problem，please take a look.
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k), and
> choose 6(b0-b5) blocks to decode other 3(b6', b7', b8') blocks. And find the
> longest common sequenece(LCS) between b6' and b6(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet is my check program output
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-15240) Erasure Coding: dirty buffer causes reconstruction block error
[
https://issues.apache.org/jira/browse/HDFS-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17067539#comment-17067539
]
HuangTao commented on HDFS-15240:
-
[~weichiu] PTAL, Thanks
> Erasure Coding: dirty buffer causes reconstruction block error
> --
>
> Key: HDFS-15240
> URL: https://issues.apache.org/jira/browse/HDFS-15240
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, erasure-coding
>Reporter: HuangTao
>Assignee: HuangTao
>Priority: Major
> Attachments: HDFS-15240.001.patch
>
>
> When read some lzo files we found some blocks were broken.
> I read back all internal blocks(b0-b8) of the block group(RS-6-3-1024k), and
> choose 6(b0-b5) blocks to decode other 3(b6', b7', b8') blocks. And find the
> longest common sequenece(LCS) between b6' and b6(b7'/b7 and b8'/b8).
> After selecting 6 blocks of the block group in combinations one time and
> iterating through all cases, I find one case that the length of LCS is the
> block length - 64KB, 64KB is just the length of ByteBuffer used by
> StripedBlockReader. So the corrupt reconstruction block is made by a dirty
> buffer.
> The following log snippet is my check program output
> {code:java}
> decode from [0, 2, 3, 4, 5, 7] -> [1, 6, 8]
> Check Block(1) first 131072 bytes longest common substring length 4
> Check Block(6) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4
> decode from [0, 2, 3, 4, 5, 6] -> [1, 7, 8]
> Check Block(1) first 131072 bytes longest common substring length 65536
> CHECK AGAIN: Block(1) all 27262976 bytes longest common substring length
> 27197440 # this one
> Check Block(7) first 131072 bytes longest common substring length 4
> Check Block(8) first 131072 bytes longest common substring length 4{code}
> Now I know the dirty buffer causes reconstruction block error, but how does
> the dirty buffer come about?
> After digging into the code and DN log, I found this following DN log is the
> root reason.
> {code:java}
> [INFO] [stripedRead-1017] : Interrupted while waiting for IO on channel
> java.nio.channels.SocketChannel[connected local=/:52586
> remote=/:50010]. 18 millis timeout left.
> [WARN] [StripedBlockReconstruction-199] : Failed to reconstruct striped
> block: BP-714356632--1519726836856:blk_-YY_3472979393
> java.lang.NullPointerException
> at
> org.apache.hadoop.hdfs.util.StripedBlockUtil.getNextCompletedStripedRead(StripedBlockUtil.java:314)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.doReadMinimumSources(StripedReader.java:308)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedReader.readMinimumSources(StripedReader.java:269)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.reconstruct(StripedBlockReconstructor.java:94)
> at
> org.apache.hadoop.hdfs.server.datanode.erasurecode.StripedBlockReconstructor.run(StripedBlockReconstructor.java:60)
> at
> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
> at
> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
> at java.base/java.lang.Thread.run(Thread.java:834) {code}
> Reading from DN may timeout(hold by a future(F)) and output the INFO log, but
> the futures that contains the future(F) is cleared,
> {code:java}
> return new StripingChunkReadResult(futures.remove(future),
> StripingChunkReadResult.CANCELLED); {code}
> futures.remove(future) cause NPE. So the EC reconstruction is failed. In the
> finally phase, the code snippet in *getStripedReader().close()*
> {code:java}
> reconstructor.freeBuffer(reader.getReadBuffer());
> reader.freeReadBuffer();
> reader.closeBlockReader(); {code}
> free buffer firstly, but the StripedBlockReader still holds the buffer and
> write it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
