[
https://issues.apache.org/jira/browse/HBASE-4518?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13142307#comment-13142307
]
Gary Helmling commented on HBASE-4518:
--------------------------------------
@Stack,
Yes, for some reason in the original code I was being a bit too "clever" and
registering the protocol handler directly. Maybe it was before all the
connecting bits had been filled in yet... In any case, the manual registration
in the original code would mean that the PingHandler would not get
re-registered if a region closed on one RS and was reopened on another. So
that is a flaw. And the test code should really be doing what we tell people
to do with endpoints, which is to configure them as coprocessors.
> TestServerCustomProtocol is flaky
> ---------------------------------
>
> Key: HBASE-4518
> URL: https://issues.apache.org/jira/browse/HBASE-4518
> Project: HBase
> Issue Type: Bug
> Components: coprocessors, test
> Affects Versions: 0.92.0
> Reporter: Gary Helmling
> Assignee: Gary Helmling
> Fix For: 0.92.0
>
> Attachments: HBASE-4518.patch,
> org.apache.hadoop.hbase.regionserver.TestServerCustomProtocol-output.txt,
> org.apache.hadoop.hbase.regionserver.TestServerCustomProtocol.txt
>
>
> TestServerCustomProtocol has been showing some intermittent failures in
> Jenkins due to what looks like region transitions.
> Here is the most recent failure:
> {noformat}
> Results :
> Failed tests:
> testRowRange(org.apache.hadoop.hbase.regionserver.TestServerCustomProtocol):
> Results should contain region
> test,bbb,1317332645939.aea9154349b9e0dc207e2e9476702763. for row 'bbb'
> {noformat}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
