[
https://issues.apache.org/jira/browse/HDFS-1921?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Matt Foley updated HDFS-1921:
-----------------------------
Attachment: hdfs-1921-2.patch
@Todd: nice tweak to the unit test. I changed the name of the subroutine to
"doTestFailedSaveNamespace", since it isn't a test case in its own right.
@Suresh:
bq. Code of thread starting logic is duplicated. It could be added to a
separate method.
Sounded right, so I implemented the suggestion, and then concluded it made the
code _more_ complex instead of better, because of the way it worked out with
the try/catch context and the management of the errorSDs list.
bq. Also continue in catch block is redundant.
The "continue"s are there for defensive coding: If someone adds statements
after the catch context, but within the loop, I believe the catch context
should go to the next loop iteration immediately.
.bq Minor: per the coding guidelines please add { } after if statements.
Done, thanks.
One more time :-)
> Save namespace can cause NN to be unable to come up on restart
> --------------------------------------------------------------
>
> Key: HDFS-1921
> URL: https://issues.apache.org/jira/browse/HDFS-1921
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.22.0, 0.23.0
> Reporter: Aaron T. Myers
> Assignee: Matt Foley
> Priority: Blocker
> Fix For: 0.22.0, 0.23.0
>
> Attachments: hdfs-1505-1-test.txt, hdfs-1921-2.patch, hdfs-1921.txt,
> hdfs1921_v23.patch, hdfs1921_v23.patch
>
>
> I discovered this in the course of trying to implement a fix for HDFS-1505.
> Per the comment for {{FSImage.saveNamespace(...)}}, the algorithm for save
> namespace proceeds in the following order:
> # rename current to lastcheckpoint.tmp for all of them,
> # save image and recreate edits for all of them,
> # rename lastcheckpoint.tmp to previous.checkpoint.
> The problem is that step 3 occurs regardless of whether or not an error
> occurs for all storage directories in step 2. Upon restart, the NN will see
> non-existent or corrupt {{current}} directories, and no
> {{lastcheckpoint.tmp}} directories, and so will conclude that the storage
> directories are not formatted.
> This issue appears to be present on both 0.22 and 0.23. This should arguably
> be a 0.22/0.23 blocker.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
