[
https://issues.apache.org/jira/browse/HBASE-10337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13871853#comment-13871853
]
Nicolas Liochon commented on HBASE-10337:
-----------------------------------------
The trunk will have the issue as well (see RpcCLient#call). As well, if the
interrupt is received during the RpcClient#writeRequest, it may be transformed
to a IOException (SocketIOWithTimeout#select), and then the connection will be
closed (among other events), which may not be what you want.
This said, we should manage interruptions correctly, in the client and in the
server code. [~jle...@virtela.net]] Do you plan to propose a patch?
> HTable.get() uninteruptible
> ---------------------------
>
> Key: HBASE-10337
> URL: https://issues.apache.org/jira/browse/HBASE-10337
> Project: HBase
> Issue Type: Bug
> Components: Client
> Affects Versions: 0.98.0, 0.94.9, 0.99.0, 0.96.1.1
> Reporter: Jonathan Leech
>
> I've got a stuck thread on HTable.get() that can't be interrupted, looks like
> its designed to be interruptible but can't be in interrupted in practice due
> to while loop.
> The offending code is in org.apache.hadoop.hbase.ipc.HBaseClient.call() line
> 981, it catches InterruptedException then goes right back to waiting due to
> the while loop.
> It looks like future versions of the client (.95+) are significantly
> different and might not have this problem... Not sure about release schedules
> etc. or if this version is still getting patched.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
