[
https://issues.apache.org/jira/browse/HBASE-8460?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Trezzo updated HBASE-8460:
--------------------------------
Attachment: HBASE-8460-trunk-v1.patch
Patch for trunk/0.95.1 attached. All replication tests pass locally.
1. The possible NPE is fixed by adding a check for null during the znode
deserialization.
2. I simply uncommented the assertion. In this failure scenario the region
server is aborted (which is a noop in the dummy region server and the test can
continue).
3. I added a check to see if the passed znode is equal to the calling region
server's znode. If so, it returns an empty map of queues.
These issues are small enough that I figured I could group them in the same
patch, but if you would like me to split them into separate JIRAs I can do that
as well.
> [replication] Fix the three TODOs in
> TestReplicationStateBasic.testReplicationQueues
> ------------------------------------------------------------------------------------
>
> Key: HBASE-8460
> URL: https://issues.apache.org/jira/browse/HBASE-8460
> Project: HBase
> Issue Type: Bug
> Components: Replication
> Reporter: Chris Trezzo
> Assignee: Chris Trezzo
> Priority: Minor
> Fix For: 0.95.1
>
> Attachments: HBASE-8460-trunk-v1.patch
>
>
> There are three TODO's in TestReplicationStateBasic.testReplicationQueues.
> 1. Possible NPE when getLogPosition is called on a Bogus HLog file.
> 2. The assertion that tested setting a log position on a bogus file was
> commented out.
> 3. When claimQueues is given the regionserver znode for itself, it will try
> to copy its own queues.
--
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
