[
https://issues.apache.org/jira/browse/HADOOP-2017?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HADOOP-2017:
--------------------------
Attachment: trsa.patch
A patch w/ more logging and thread dumping to better help what is going on, and
a mechanism that notices moved regions sooner.
{code}
HADOOP-2017 TestRegionServerAbort failure in patch build #903 and nightly #266
Notice moved META regions sooner. Also added more logging and
thread dumping once a minute when test starts to take too long
so can see where we are hung (if we are hung).
M src/contrib/hbase/src/test/org/apache/hadoop/hbase/TestHStoreFile.java
Inherit from HBaseTestCase.
M src/contrib/hbase/src/test/org/apache/hadoop/hbase/HBaseClusterTestCase.java
(threadDumpingJoin): Added.
M src/contrib/hbase/src/test/org/apache/hadoop/hbase/TestRegionServerAbort.java
Run verification in its own thread so can concurrently thread dump if
test is going on too long.
M src/contrib/hbase/src/test/org/apache/hadoop/hbase/DFSAbort.java
Moved join up into parent class.
M src/contrib/hbase/src/java/org/apache/hadoop/hbase/Chore.java
Remove unused import.
M src/contrib/hbase/src/java/org/apache/hadoop/hbase/HMaster.java
(MetaRegion.toString): Added.
Added logging around assignment checking and log split.
(MetaRegion.compareTo): Add consideration of server address.
(numberOfMetaRegions, metaRegionsToScan, onlineMetaRegions):
Put declaration and assignment together and made final.
(scanOneMetaRegion): If the region is no longer in onlineMetaRegions,
give up trying to scan.
(unassignRootRegion): Added (Not yet finished).
{code}
> [hbase] TestRegionServerAbort failure in patch build #903 and nightly #266
> --------------------------------------------------------------------------
>
> Key: HADOOP-2017
> URL: https://issues.apache.org/jira/browse/HADOOP-2017
> Project: Hadoop
> Issue Type: Bug
> Components: contrib/hbase
> Reporter: stack
> Priority: Minor
> Fix For: 0.15.0
>
> Attachments: trsa.patch
>
>
> In patch build #903, the metascanner keeps trying to go to the downed server
> even though onlineMetaRegions has been updated w/ new location and then the
> metascanner just goes away (or hangs).
> In nightly build #266, its a similar scenario only the remaining region
> servers decide to shut down because they haven't been able to reach the
> master in 7 seconds.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
