[
https://issues.apache.org/jira/browse/HBASE-12747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrew Purtell updated HBASE-12747:
-----------------------------------
Resolution: Fixed
Hadoop Flags: Reviewed
Status: Resolved (was: Patch Available)
Pushed to 0.98+.
I tested setting new 'surefire.Xmx' and 'failsafe.Xmx' properties for "mvn
test" and "mvn verify", respectively: "3g" launches JVMs with expected heap
size, "bogus" causes JVM fork failures as expected. New default lets
IntegrationTestMTTR pass when run with "mvn verify" (tested with 0.98 and 1.0).
Thanks for the patch [~abhishek.chouhan]
> IntegrationTestMTTR will OOME if launched with mvn verify
> ---------------------------------------------------------
>
> Key: HBASE-12747
> URL: https://issues.apache.org/jira/browse/HBASE-12747
> Project: HBase
> Issue Type: Bug
> Affects Versions: 0.98.9
> Reporter: Andrew Purtell
> Assignee: Abhishek Singh Chouhan
> Priority: Minor
> Fix For: 2.0.0, 1.0.1, 1.1.0, 0.98.11
>
> Attachments: HBASE-12747-v1.patch, HBASE-12747.patch,
> org.apache.hadoop.hbase.mttr.IntegrationTestMTTR-output.txt.gz
>
>
> IntegrationTestMTRR will OOME if launched like:
> {noformat}
> cd hbase-it
> mvn verify -Dit.test=IntegrationTestMTTR
> {noformat}
> Linux environment, 7u67.
> Looks like we should bump the heap on the failsafe argline in the POM.
> {noformat}
> 2014-12-22 11:24:07,725 ERROR
> [B.DefaultRpcServer.handler=2,queue=0,port=55672] ipc.RpcServer(2067):
> Unexpected throwable o
> bject
> java.lang.OutOfMemoryError: Java heap space
> at
> org.apache.hadoop.hbase.regionserver.MemStoreLAB$Chunk.init(MemStoreLAB.java:246)
> at
> org.apache.hadoop.hbase.regionserver.MemStoreLAB.getOrMakeChunk(MemStoreLAB.java:196)
> at
> org.apache.hadoop.hbase.regionserver.MemStoreLAB.allocateBytes(MemStoreLAB.java:114)
> at
> org.apache.hadoop.hbase.regionserver.MemStore.maybeCloneWithAllocator(MemStore.java:274)
> at
> org.apache.hadoop.hbase.regionserver.MemStore.add(MemStore.java:229)
> at org.apache.hadoop.hbase.regionserver.HStore.add(HStore.java:576)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.applyFamilyMapToMemstore(HRegion.java:3084)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.doMiniBatchMutation(HRegion.java:2517)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.batchMutate(HRegion.java:2284)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.batchMutate(HRegion.java:2239)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.batchMutate(HRegion.java:2243)
> at
> org.apache.hadoop.hbase.regionserver.HRegionServer.doBatchOp(HRegionServer.java:4482)
> at
> org.apache.hadoop.hbase.regionserver.HRegionServer.doNonAtomicRegionMutation(HRegionServer.java:3665)
> at
> org.apache.hadoop.hbase.regionserver.HRegionServer.multi(HRegionServer.java:3554)
> {noformat}
> Another minor issue: After taking the OOME, the test executor will linger
> indefinitely as a zombie.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
