[
https://issues.apache.org/jira/browse/HDFS-1262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12882160#action_12882160
]
Todd Lipcon commented on HDFS-1262:
-----------------------------------
Oof, what a hack that would be :) Not to say we shouldn't do that over in HBase
in the short term, but I agree let's treat this as a serious bug in 0.20-append
and try to fix it on the HDFS side unless we really can't think of any
implementable solutions.
Can you think of a problem with the abandonBlock() solution? My thinking is
that we'd check if the block is the last block of a file under construction by
the abandoning client, and if so, reassign lease to NN_Recovery and initiate
block synchronization from the NN as if the lease were lost. It may not be
necessary to go through the whole recovery process, but it will be safer in
case the client half set up a pipeline before failing, or somesuch.
> Failed pipeline creation during append leaves lease hanging on NN
> -----------------------------------------------------------------
>
> Key: HDFS-1262
> URL: https://issues.apache.org/jira/browse/HDFS-1262
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs client, name-node
> Affects Versions: 0.20-append
> Reporter: Todd Lipcon
> Priority: Critical
> Fix For: 0.20-append
>
>
> Ryan Rawson came upon this nasty bug in HBase cluster testing. What happened
> was the following:
> 1) File's original writer died
> 2) Recovery client tried to open file for append - looped for a minute or so
> until soft lease expired, then append call initiated recovery
> 3) Recovery completed successfully
> 4) Recovery client calls append again, which succeeds on the NN
> 5) For some reason, the block recovery that happens at the start of append
> pipeline creation failed on all datanodes 6 times, causing the append() call
> to throw an exception back to HBase master. HBase assumed the file wasn't
> open and put it back on a queue to try later
> 6) Some time later, it tried append again, but the lease was still assigned
> to the same DFS client, so it wasn't able to recover.
> The recovery failure in step 5 is a separate issue, but the problem for this
> JIRA is that the NN can think it failed to open a file for append when the NN
> thinks the writer holds a lease. Since the writer keeps renewing its lease,
> recovery never happens, and no one can open or recover the file until the DFS
> client shuts down.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
