[
https://issues.apache.org/jira/browse/HBASE-14650?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14980147#comment-14980147
]
Heng Chen commented on HBASE-14650:
-----------------------------------
The reason is that:
After we do split, we split again immediately. The reference files are not
removed, so the secondary split will failed.
And we can see logs like below.
{code}
2015-10-29 17:23:15,458 INFO [PriorityRpcServer.handler=1,queue=1,port=49462]
regionserver.RSRpcServices(1845): Splitting
TestNamespaceAuditor_regiontest:table1,,1446110595269.81e71c20136ae82ac59aeac72a8d67b4.
2015-10-29 17:23:15,459 DEBUG [PriorityRpcServer.handler=1,queue=1,port=49462]
regionserver.HStore(1882): Cannot split region due to reference files being
there
{code}
So the main thread hangs at the secondary countdown latch.
{code}
ADMIN.split(tableOne, getSplitKey(actualRegion.getRegionInfo().getStartKey(),
actualRegion.getRegionInfo().getEndKey()));
observer.postSplit.await();
{code}
To fix this, we should do compaction after first split. And testcase will
pass.
> Reenable TestNamespaceAuditor
> -----------------------------
>
> Key: HBASE-14650
> URL: https://issues.apache.org/jira/browse/HBASE-14650
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Attachments: HBASE-14650.patch
>
>
> Was disabled over in HBASE-14569 because flakey. Reenable when fixed.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
