[
https://issues.apache.org/jira/browse/HBASE-8583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13664568#comment-13664568
]
Enis Soztutar commented on HBASE-8583:
--------------------------------------
Did you test this with:
- Running as a unit test
- Running from ITDriver against (pseudo) dist cluster
- Running from command line against (pseudo) dist cluster (you may need to
call IntegrationTestingUtility.setUseDistributedCluster(conf);)
{code}
+ * Short string representation without param info because param itself
could huge depends on the
{code}
could *be* huge?
While we are at it, can you please change toShortString() to use StringBuilder?
Also (call.size > 512) ? call.toShortString() : call.toString() can go inside
toString().
>From earlier offline discussions with Jeff, I think the problem for this IT
>being run as a UT was that we CM will kill the RS, but on kill, we do not
>close the WAL, thus LogSyncer daemon thread will just live forever and
>continue to output logs. The root cause of the problem is actually that we try
>to simulate an RS kill by calling HRS.kill() instead of abort, but since we do
>not quit the process, daemon threads are not killed. I agree that if we are
>killed, we should not try to close wal, but on the other hand, if we do not,
>nobody is closing the logsycner thread.
Lets also get this in for 0.94 after trunk patch is committed.
> Create a new IntegrationTestBigLinkedListWithChaosMonkey
> --------------------------------------------------------
>
> Key: HBASE-8583
> URL: https://issues.apache.org/jira/browse/HBASE-8583
> Project: HBase
> Issue Type: Sub-task
> Components: test
> Reporter: Jeffrey Zhong
> Assignee: Jeffrey Zhong
> Fix For: 0.95.1
>
> Attachments: hbase-8583.patch, hbase-8583-v1.patch,
> hbase-8583-v2.patch
>
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
