[
https://issues.apache.org/jira/browse/HDFS-5405?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrew Wang updated HDFS-5405:
------------------------------
Attachment: hdfs-5405-1.patch
Patch attached, just FSNamesystem fixes. Basically the idea is to move anything
that might throw an exception and not set the retry cache state before
{{RetryCache.waitForCompletion}}.
> Fix possible RetryCache hang for caching RPC handlers in FSNamesystem
> ---------------------------------------------------------------------
>
> Key: HDFS-5405
> URL: https://issues.apache.org/jira/browse/HDFS-5405
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: namenode
> Affects Versions: HDFS-4949
> Reporter: Andrew Wang
> Assignee: Andrew Wang
> Attachments: hdfs-5405-1.patch
>
>
> Related to the fix in HDFS-5299, we need to make sure that uncaught
> exceptions happen before {{RetryCache.waitForCompletion}}, or badly timed
> failovers can lead to a perma-hang.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
