[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15396704#comment-15396704 ] Arpit Agarwal commented on HDFS-9259: - Thanks for the quick review Mingliang. > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15396693#comment-15396693 ] Mingliang Liu commented on HDFS-9259: - Thanks [~arpitagarwal] for adding release note. It looks pretty good to me. > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[
https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14976578#comment-14976578
]
Hadoop QA commented on HDFS-9259:
-
\\
\\
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:red}-1{color} | pre-patch | 20m 13s | Pre-patch trunk has 1 extant
Findbugs (version 3.0.0) warnings. |
| {color:green}+1{color} | @author | 0m 0s | The patch does not contain any
@author tags. |
| {color:green}+1{color} | tests included | 0m 0s | The patch appears to
include 1 new or modified test files. |
| {color:green}+1{color} | javac | 8m 4s | There were no new javac warning
messages. |
| {color:green}+1{color} | javadoc | 10m 33s | There were no new javadoc
warning messages. |
| {color:green}+1{color} | release audit | 0m 25s | The applied patch does
not increase the total number of release audit warnings. |
| {color:green}+1{color} | checkstyle | 2m 43s | There were no new checkstyle
issues. |
| {color:green}+1{color} | whitespace | 0m 0s | The patch has no lines that
end in whitespace. |
| {color:green}+1{color} | install | 1m 33s | mvn install still works. |
| {color:green}+1{color} | eclipse:eclipse | 0m 34s | The patch built with
eclipse:eclipse. |
| {color:green}+1{color} | findbugs | 4m 29s | The patch does not introduce
any new Findbugs (version 3.0.0) warnings. |
| {color:green}+1{color} | native | 3m 12s | Pre-build of native portion |
| {color:red}-1{color} | hdfs tests | 49m 23s | Tests failed in hadoop-hdfs. |
| {color:green}+1{color} | hdfs tests | 0m 32s | Tests passed in
hadoop-hdfs-client. |
| | | 101m 44s | |
\\
\\
|| Reason || Tests ||
| Failed unit tests |
hadoop.hdfs.server.balancer.TestBalancerWithMultipleNameNodes |
\\
\\
|| Subsystem || Report/Notes ||
| Patch URL |
http://issues.apache.org/jira/secure/attachment/12768842/HDFS-9259.001.patch |
| Optional Tests | javadoc javac unit findbugs checkstyle |
| git revision | trunk / bcb2386 |
| Pre-patch Findbugs warnings |
https://builds.apache.org/job/PreCommit-HDFS-Build/13224/artifact/patchprocess/trunkFindbugsWarningshadoop-hdfs.html
|
| hadoop-hdfs test log |
https://builds.apache.org/job/PreCommit-HDFS-Build/13224/artifact/patchprocess/testrun_hadoop-hdfs.txt
|
| hadoop-hdfs-client test log |
https://builds.apache.org/job/PreCommit-HDFS-Build/13224/artifact/patchprocess/testrun_hadoop-hdfs-client.txt
|
| Test Results |
https://builds.apache.org/job/PreCommit-HDFS-Build/13224/testReport/ |
| Java | 1.7.0_55 |
| uname | Linux asf904.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP
PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux |
| Console output |
https://builds.apache.org/job/PreCommit-HDFS-Build/13224/console |
This message was automatically generated.
> Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
> --
>
> Key: HDFS-9259
> URL: https://issues.apache.org/jira/browse/HDFS-9259
> Project: Hadoop HDFS
> Issue Type: Improvement
>Reporter: Ming Ma
>Assignee: Mingliang Liu
> Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch
>
>
> We recently found that cross-DC hdfs write could be really slow. Further
> investigation identified that is due to SendBufferSize and ReceiveBufferSize
> used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file
> across DC with different SendBufferSize and ReceiveBufferSize values. The
> results showed that c much faster than b; b is faster than a.
> a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting).
> b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning).
> c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both)
> HDFS-8829 has enabled scenario b. We would like to enable scenario c by
> making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He
> Tianyi] [~kanaka] [~vinayrpet].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14976942#comment-14976942 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-Hdfs-trunk-Java8 #543 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk-Java8/543/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14976871#comment-14976871 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-trunk-Commit #8715 (See [https://builds.apache.org/job/Hadoop-trunk-Commit/8715/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14977044#comment-14977044 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-Yarn-trunk-Java8 #604 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk-Java8/604/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14977012#comment-14977012 ] Hudson commented on HDFS-9259: -- SUCCESS: Integrated in Hadoop-Yarn-trunk #1327 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk/1327/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14977037#comment-14977037 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-Mapreduce-trunk-Java8 #591 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk-Java8/591/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14977312#comment-14977312 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-Mapreduce-trunk #2535 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/2535/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14977338#comment-14977338 ] Hudson commented on HDFS-9259: -- FAILURE: Integrated in Hadoop-Hdfs-trunk #2481 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/2481/]) HDFS-9259. Make SO_SNDBUF size configurable at DFSClient side for hdfs (mingma: rev aa09880ab85f3c35c12373976e7b03f3140b65c8) * hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/DfsClientConf.java * hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientSocketSize.java * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/HdfsClientConfigKeys.java * hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml * hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DataStreamer.java > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Fix For: 2.8.0 > > Attachments: HDFS-9259.000.patch, HDFS-9259.001.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14975177#comment-14975177 ] Ming Ma commented on HDFS-9259: --- Thanks [~liuml07]! The patch looks good. Just some format nits in TestDFSClientSocketSize: testAutoTuningSendBufferSize's assertTrue statement's indent is off; createSocket's createSocketForPipeline statement can put the last two lines into one line. > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma >Assignee: Mingliang Liu > Attachments: HDFS-9259.000.patch > > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file > across DC with different SendBufferSize and ReceiveBufferSize values. The > results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c by > making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He > Tianyi] [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[
https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14973065#comment-14973065
]
Hadoop QA commented on HDFS-9259:
-
\\
\\
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:red}-1{color} | pre-patch | 22m 42s | Pre-patch trunk has 1 extant
Findbugs (version 3.0.0) warnings. |
| {color:green}+1{color} | @author | 0m 0s | The patch does not contain any
@author tags. |
| {color:green}+1{color} | tests included | 0m 0s | The patch appears to
include 1 new or modified test files. |
| {color:green}+1{color} | javac | 8m 48s | There were no new javac warning
messages. |
| {color:green}+1{color} | javadoc | 11m 15s | There were no new javadoc
warning messages. |
| {color:green}+1{color} | release audit | 0m 25s | The applied patch does
not increase the total number of release audit warnings. |
| {color:green}+1{color} | checkstyle | 2m 55s | There were no new checkstyle
issues. |
| {color:green}+1{color} | whitespace | 0m 0s | The patch has no lines that
end in whitespace. |
| {color:green}+1{color} | install | 1m 56s | mvn install still works. |
| {color:green}+1{color} | eclipse:eclipse | 0m 40s | The patch built with
eclipse:eclipse. |
| {color:green}+1{color} | findbugs | 5m 4s | The patch does not introduce
any new Findbugs (version 3.0.0) warnings. |
| {color:green}+1{color} | native | 3m 34s | Pre-build of native portion |
| {color:red}-1{color} | hdfs tests | 67m 2s | Tests failed in hadoop-hdfs. |
| {color:green}+1{color} | hdfs tests | 0m 32s | Tests passed in
hadoop-hdfs-client. |
| | | 124m 56s | |
\\
\\
|| Reason || Tests ||
| Failed unit tests | hadoop.tools.TestHdfsConfigFields |
| | hadoop.hdfs.TestRecoverStripedFile |
| | hadoop.hdfs.server.namenode.TestFileTruncate |
| | hadoop.hdfs.server.blockmanagement.TestUnderReplicatedBlocks |
\\
\\
|| Subsystem || Report/Notes ||
| Patch URL |
http://issues.apache.org/jira/secure/attachment/12768573/HDFS-9259.000.patch |
| Optional Tests | javadoc javac unit findbugs checkstyle |
| git revision | trunk / 446212a |
| Pre-patch Findbugs warnings |
https://builds.apache.org/job/PreCommit-HDFS-Build/13184/artifact/patchprocess/trunkFindbugsWarningshadoop-hdfs.html
|
| hadoop-hdfs test log |
https://builds.apache.org/job/PreCommit-HDFS-Build/13184/artifact/patchprocess/testrun_hadoop-hdfs.txt
|
| hadoop-hdfs-client test log |
https://builds.apache.org/job/PreCommit-HDFS-Build/13184/artifact/patchprocess/testrun_hadoop-hdfs-client.txt
|
| Test Results |
https://builds.apache.org/job/PreCommit-HDFS-Build/13184/testReport/ |
| Java | 1.7.0_55 |
| uname | Linux asf909.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP
PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux |
| Console output |
https://builds.apache.org/job/PreCommit-HDFS-Build/13184/console |
This message was automatically generated.
> Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
> --
>
> Key: HDFS-9259
> URL: https://issues.apache.org/jira/browse/HDFS-9259
> Project: Hadoop HDFS
> Issue Type: Improvement
>Reporter: Ming Ma
>Assignee: Mingliang Liu
> Attachments: HDFS-9259.000.patch
>
>
> We recently found that cross-DC hdfs write could be really slow. Further
> investigation identified that is due to SendBufferSize and ReceiveBufferSize
> used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file
> across DC with different SendBufferSize and ReceiveBufferSize values. The
> results showed that c much faster than b; b is faster than a.
> a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting).
> b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning).
> c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both)
> HDFS-8829 has enabled scenario b. We would like to enable scenario c by
> making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He
> Tianyi] [~kanaka] [~vinayrpet].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[
https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14973448#comment-14973448
]
Mingliang Liu commented on HDFS-9259:
-
The failing test {{TestHdfsConfigFields}} is tracked by [HDFS-9304].
> Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
> --
>
> Key: HDFS-9259
> URL: https://issues.apache.org/jira/browse/HDFS-9259
> Project: Hadoop HDFS
> Issue Type: Improvement
>Reporter: Ming Ma
>Assignee: Mingliang Liu
> Attachments: HDFS-9259.000.patch
>
>
> We recently found that cross-DC hdfs write could be really slow. Further
> investigation identified that is due to SendBufferSize and ReceiveBufferSize
> used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file
> across DC with different SendBufferSize and ReceiveBufferSize values. The
> results showed that c much faster than b; b is faster than a.
> a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting).
> b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning).
> c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both)
> HDFS-8829 has enabled scenario b. We would like to enable scenario c by
> making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He
> Tianyi] [~kanaka] [~vinayrpet].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[
https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14964250#comment-14964250
]
Colin Patrick McCabe commented on HDFS-9259:
That seems reasonable to me. It sounds like we could have some new client
configs for this... with {{client}} somewhere in the config key name.
> Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
> --
>
> Key: HDFS-9259
> URL: https://issues.apache.org/jira/browse/HDFS-9259
> Project: Hadoop HDFS
> Issue Type: Improvement
>Reporter: Ming Ma
>Assignee: Mingliang Liu
>
> We recently found that cross-DC hdfs write could be really slow. Further
> investigation identified that is due to SendBufferSize and ReceiveBufferSize
> used for hdfs write. The test ran "hadoop -fs -copyFromLocal" of a 256MB file
> across DC with different SendBufferSize and ReceiveBufferSize values. The
> results showed that c much faster than b; b is faster than a.
> a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting).
> b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning).
> c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both)
> HDFS-8829 has enabled scenario b. We would like to enable scenario c by
> making SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He
> Tianyi] [~kanaka] [~vinayrpet].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (HDFS-9259) Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario
[ https://issues.apache.org/jira/browse/HDFS-9259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14961091#comment-14961091 ] Mingliang Liu commented on HDFS-9259: - Hi [~mingma], can I work on this, if we reach consensus on the issue itself? > Make SO_SNDBUF size configurable at DFSClient side for hdfs write scenario > -- > > Key: HDFS-9259 > URL: https://issues.apache.org/jira/browse/HDFS-9259 > Project: Hadoop HDFS > Issue Type: Improvement >Reporter: Ming Ma > > We recently found that cross-DC hdfs write could be really slow. Further > investigation identified that is due to SendBufferSize and ReceiveBufferSize > used for hdfs write. The test is to do "hadoop -fs -copyFromLocal" of a 256MB > file across DC with different SendBufferSize and ReceiveBufferSize values. > The results showed that c much faster than b; b is faster than a. > a. SendBufferSize=128k, ReceiveBufferSize=128k (hdfs default setting). > b. SendBufferSize=128K, ReceiveBufferSize=not set(TCP auto tuning). > c. SendBufferSize=not set, ReceiveBufferSize=not set(TCP auto tuning for both) > HDFS-8829 has enabled scenario b. We would like to enable scenario c to make > SendBufferSize configurable at DFSClient side. Cc: [~cmccabe] [~He Tianyi] > [~kanaka] [~vinayrpet]. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
