[
https://issues.apache.org/jira/browse/HADOOP-6762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12876081#action_12876081
]
Todd Lipcon commented on HADOOP-6762:
-------------------------------------
bq. btw, is it appropriate to delete the very old patches as they aren't really
relevant? or just leave them?
Usually just leave them - makes it easier to follow the conversation and see
how the patch evolved.
bq. (which there might a timeout on the actual out.write call already--i need
to double-check)
Looks like there is one - a timeout is set when NetUtils.getOutputStream is
called on the socket in setupIOStreams(). So I don't think we need a separate
timeout in awaiting the send param.
I think the patch is good. I ran it with my test case since Friday night and
didn't see any RPC hangs. The test case eventually failed with "Too many open
files" but I think it's some other bug/socket leak in DFS code, not IPC. Mark
Patch Available to swing it through Hudson?
> exception while doing RPC I/O closes channel
> --------------------------------------------
>
> Key: HADOOP-6762
> URL: https://issues.apache.org/jira/browse/HADOOP-6762
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 0.20.2
> Reporter: sam rash
> Assignee: sam rash
> Attachments: hadoop-6762-1.txt, hadoop-6762-2.txt, hadoop-6762-3.txt,
> hadoop-6762-4.txt, hadoop-6762-6.txt, hadoop-6762-7.txt
>
>
> If a single process creates two unique fileSystems to the same NN using
> FileSystem.newInstance(), and one of them issues a close(), the leasechecker
> thread is interrupted. This interrupt races with the rpc namenode.renew()
> and can cause a ClosedByInterruptException. This closes the underlying
> channel and the other filesystem, sharing the connection will get errors.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
