[
https://issues.apache.org/jira/browse/HBASE-19988?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16363378#comment-16363378
]
Appy commented on HBASE-19988:
------------------------------
Sorry, i don't have time to dig in and come up with a better understanding of
handling InterruptedException when processing requests.
In this case, since IE was already being converted to IIOE, that means any
other operation would have been handling it like IOException, which means
cancel the operation. Going by that logic, and status quo bias (that it's
already IIOE), i think it might be fine to do this.
However, I think it'll be better to handle it as part of IOException by doing
{code}
if (isAtomic() or ioe instanceof IIOE) { throw ioe; }
{code}
because it'll log a good warning.
Maybe move TimeoutIOException there too.
Currently the comment says "// We will retry when other exceptions, but we
should stop if we timeout ."
Should be updated with reasons why we break out for each type. Let's not leave
things in more dismay for future onlookers (why these two? why not others? etc
etc). They shouldn't have to spend the time we already did, else our effort is
wasted.
> HRegion#lockRowsAndBuildMiniBatch() is too chatty when interrupted while
> waiting for a row lock
> -----------------------------------------------------------------------------------------------
>
> Key: HBASE-19988
> URL: https://issues.apache.org/jira/browse/HBASE-19988
> Project: HBase
> Issue Type: Improvement
> Components: amv2
> Affects Versions: 2.0.0-beta-1
> Reporter: Umesh Agashe
> Assignee: Umesh Agashe
> Priority: Minor
> Fix For: 2.0.0-beta-2
>
> Attachments: hbase-19988.master.001.patch,
> hbase-19988.master.001.patch
>
>
> See HBASE-19970, TestHRegionWithInMemoryFlush created 4.2g log file.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
