[
https://issues.apache.org/jira/browse/HDFS-1142?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12875298#action_12875298
]
Hairong Kuang commented on HDFS-1142:
-------------------------------------
Interesting discussion...
As long as I understand, once lease recovery starts, the first thing that each
replica datanode does is kill the block receiver. So the old client will fail
since the pipleline is completely torn down.The idea is that no writer is
allowed, i.e., no modification to the file is allowed from any client, while
the file's lease recovery is in progress. So in this sense, it would be nice to
reassign the lease to NN but still maintain the soft lease semantics that
Konstantin described.
> Lease recovery doesn't reassign lease when triggered by append()
> ----------------------------------------------------------------
>
> Key: HDFS-1142
> URL: https://issues.apache.org/jira/browse/HDFS-1142
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: name-node
> Affects Versions: 0.21.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1142.txt, hdfs-1142.txt
>
>
> If a soft lease has expired and another writer calls append(), it triggers
> lease recovery but doesn't reassign the lease to a new owner. Therefore, the
> old writer can continue to allocate new blocks, try to steal back the lease,
> etc. This is for the testRecoveryOnBlockBoundary case of HDFS-1139
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
