[
https://issues.apache.org/jira/browse/HBASE-8164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13618124#comment-13618124
]
ramkrishna.s.vasudevan commented on HBASE-8164:
-----------------------------------------------
The reason for failure should be
In getSplittableRegion()
{code}
regions = getHBaseCluster().getRegions(tableName);
{code}
The regions are obtained from the RS memory (the online regions in that RS).
Now as per the logs in the failed test run, it seems that the region for which
the split failed is the one that just got newly created as part of its parent's
split.
Now as this region is added to online region list we tend to check if that new
region can be splitted.
Ideally when we check if region.checkSplit() in getSplittableRegions
{code}
if (!splitPolicy.shouldSplit()) {
return null;
}
{code}
The shouldSplit will return false here. This is because there are no store
files for this region and also forceSplit is false.
You can see that the compaction has not happened for the this region due to the
same reason.
But when we issue admin.split() the above shouldSplit is returning true because
forceSplit is true now. And hence it goes upto getSplitPoint() and fails
there.
{code}
2013-03-30 04:43:50,481 INFO [PRI IPC Server handler 8 on 56813]
regionserver.HRegionServer(3551): Compacting
testTableReadLock,,1364618626029.be3c0267dd27d6b83ffd4b53832987e2.
2013-03-30 04:43:50,481 DEBUG [PRI IPC Server handler 8 on 56813]
compactions.DefaultCompactionPolicy(222): Not compacting files because we only
have 0 files ready for compaction. Need 3 to initiate.
2013-03-30 04:43:50,481 DEBUG [PRI IPC Server handler 8 on 56813]
regionserver.CompactSplitThread(281): Not compacting
testTableReadLock,,1364618626029.be3c0267dd27d6b83ffd4b53832987e2. because
compaction request was cancelled
{code}
Pls do correct me if am wrong here.
> TestTableLockManager fails intermittently in trunk builds
> ---------------------------------------------------------
>
> Key: HBASE-8164
> URL: https://issues.apache.org/jira/browse/HBASE-8164
> Project: HBase
> Issue Type: Bug
> Reporter: Ted Yu
> Assignee: Ted Yu
> Fix For: 0.95.0, 0.98.0
>
> Attachments: 8164-addendum.txt, 8164-v2.txt, 8164-v3.txt, 8164-v4.txt
>
>
> In build #3979:
> testTableReadLock(org.apache.hadoop.hbase.master.TestTableLockManager): test
> timed out after 600000 milliseconds
--
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
