You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "stack (JIRA)" <ji...@apache.org> on 2011/04/30 01:29:03 UTC
[jira] [Commented] (HBASE-3832) Failing
TestMasterFailover.testMasterFailoverWithMockedRITOnDeadRS up on jenkins
[ https://issues.apache.org/jira/browse/HBASE-3832?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13027235#comment-13027235 ]
stack commented on HBASE-3832:
------------------------------
Committed patch.
> Failing TestMasterFailover.testMasterFailoverWithMockedRITOnDeadRS up on jenkins
> --------------------------------------------------------------------------------
>
> Key: HBASE-3832
> URL: https://issues.apache.org/jira/browse/HBASE-3832
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Attachments: 3832.txt
>
>
> Root region is stuck in RIT.
> Seems to be because of this:
> 3316 2011-04-29 05:53:11,941 WARN [Thread-642-EventThread] master.AssignmentManager(518): Received OPENED for region 70236052/-ROOT- from server vesta.apache.org,57336,1304056370834 but region was in the state null and not in expected PENDING_OPEN or OPENING states
> Later I see this:
> 3334 2011-04-29 05:53:12,014 DEBUG [Master:0;vesta.apache.org,36450,1304056384388] master.AssignmentManager(260): Found REGION => {NAME => '-ROOT-,,0', STARTKEY => '', ENDKEY => '', ENCODED => 70236052, TABLE => {{NAME => '-ROOT-', IS_ROOT => 'true', IS_META => 'true', FAMILIES => [{NAME => 'info', BLOOMFILTER => 'NONE', REPLICATION_SCOPE => '0', COMPRESSION => 'NONE', VERSIONS => '10', TTL => '2147483647', BLOCKSIZE => '8192', IN_MEMORY => 'true', BLOCKCACHE => 'true'}]}}=vesta.apache.org,57336,1304056370834 in RITs
> The former makes it so we don't clear a successfully opened -ROOT- from RIT so we get the second line and then the test fails with:
> 5192 2011-04-29 05:55:42,181 DEBUG [Thread-642] zookeeper.ZKAssign(815): ZK RIT -> 70236052
> printed over and over again.
> I don't get why the data is null in the zk when RS has updated it a couple of times. I see we do a double regionOnline in master code. This clears in-memory master state. I don't think this it but will commit this and more logging to help w/ the debug.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira