[
https://issues.apache.org/jira/browse/HADOOP-1134?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Raghu Angadi updated HADOOP-1134:
---------------------------------
Attachment: HADOOP-1134-01.patch
Finally attaching the patch. This patch needs patch from HADOOP-1597.
All the tests pass. The patch has been tested quite a bit. Upgrade also ran on
a large cluster and tested well.
Following files are added under dfs/ : DataChecksum.java BlockCrcUpgrade.java
ChecksumDistributedFileSystem.java
There is an observed slow down of may be 5-10% (smaller the cluster, larger
the difference) in sort benchmarks. I am pretty sure it is related buffering
changes made in this patch. I just did not get enough time to investigate them
well. I am pretty sure this can be fixed. will be looking into next. Most of
the difference also shows up in randomWriter.
Will write up more on Upgrade guide for cluster admins. It will include details
on how to monitor the progress and how to handle some error situations.
Feedback is welcome.
> Block level CRCs in HDFS
> ------------------------
>
> Key: HADOOP-1134
> URL: https://issues.apache.org/jira/browse/HADOOP-1134
> Project: Hadoop
> Issue Type: New Feature
> Components: dfs
> Reporter: Raghu Angadi
> Assignee: Raghu Angadi
> Attachments: BlockLevelCrc-07032007.patch,
> BlockLevelCrc-07052007.patch, BlockLevelCrc-07062007.patch,
> BlockLevelCrc-07102007.patch, BlockLevelCrc-07122007.patch,
> DfsBlockCrcDesign-05305007.htm, HADOOP-1134-01.patch, readBuffer.java
>
>
> Currently CRCs are handled at FileSystem level and are transparent to core
> HDFS. See recent improvement HADOOP-928 ( that can add checksums to a given
> filesystem ) regd more about it. Though this served us well there a few
> disadvantages :
> 1) This doubles namespace in HDFS ( or other filesystem implementations ). In
> many cases, it nearly doubles the number of blocks. Taking namenode out of
> CRCs would nearly double namespace performance both in terms of CPU and
> memory.
> 2) Since CRCs are transparent to HDFS, it can not actively detect corrupted
> blocks. With block level CRCs, Datanode can periodically verify the checksums
> and report corruptions to namnode such that name replicas can be created.
> We propose to have CRCs maintained for all HDFS data in much the same way as
> in GFS. I will update the jira with detailed requirements and design. This
> will include same guarantees provided by current implementation and will
> include a upgrade of current data.
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
