[
https://issues.apache.org/jira/browse/PHOENIX-6274?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17252486#comment-17252486
]
Viraj Jasani commented on PHOENIX-6274:
---------------------------------------
Thanks [~stoty]. Agree on disabling balancer, it doesn't harm and is quite safe
option. However, without this patch, as of today, we are not successfully
splitting the table and hence we are not testing snapshot creation on multi
region in reality. However, with the patch, we will start splitting table and
there are chances that balancer can kick in and move a region, hence,
definitely good to consider turning off the balancer. Let me add it in PR and
upload updated patches.
Thanks for the suggestion!
> Flaky test TableSnapshotReadsMapReduceIT.testMapReduceSnapshotsMultiRegion
> --------------------------------------------------------------------------
>
> Key: PHOENIX-6274
> URL: https://issues.apache.org/jira/browse/PHOENIX-6274
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 5.0.0, 4.15.0
> Reporter: Viraj Jasani
> Assignee: Viraj Jasani
> Priority: Major
> Fix For: 5.1.0, 4.16.0
>
> Attachments: PHOENIX-6274.4.x.000.patch, PHOENIX-6274.master.000.patch
>
>
> Test TableSnapshotReadsMapReduceIT.testMapReduceSnapshotsMultiRegion
> sometimes fails to delete snapshot due to snapshot corruption.
> {code:java}
> ERROR [MASTER_SNAPSHOT_OPERATIONS-master/e09b7a102100:0-0]
> org.apache.hadoop.hbase.master.snapshot.MasterSnapshotVerifier(187): Regions
> moved during the snapshot '{ ss=FOO table=N000006 type=FLUSH }'. expected=2
> snapshotted=3.ERROR [MASTER_SNAPSHOT_OPERATIONS-master/e09b7a102100:0-0]
> org.apache.hadoop.hbase.master.snapshot.MasterSnapshotVerifier(187): Regions
> moved during the snapshot '{ ss=FOO table=N000006 type=FLUSH }'. expected=2
> snapshotted=3.2020-12-18 01:18:45,532 ERROR
> [MASTER_SNAPSHOT_OPERATIONS-master/e09b7a102100:0-0]
> org.apache.hadoop.hbase.master.snapshot.TakeSnapshotHandler(222): Failed
> taking snapshot { ss=FOO table=N000006 type=FLUSH } due to exception:Regions
> moved during the snapshot '{ ss=FOO table=N000006 type=FLUSH }'. expected=2
> snapshotted=3.org.apache.hadoop.hbase.snapshot.CorruptedSnapshotException:
> Regions moved during the snapshot '{ ss=FOO table=N000006 type=FLUSH }'.
> expected=2 snapshotted=3. at
> org.apache.hadoop.hbase.master.snapshot.MasterSnapshotVerifier.verifyRegions(MasterSnapshotVerifier.java:205)
> at
> org.apache.hadoop.hbase.master.snapshot.MasterSnapshotVerifier.verifySnapshot(MasterSnapshotVerifier.java:119)
> at
> org.apache.hadoop.hbase.master.snapshot.TakeSnapshotHandler.process(TakeSnapshotHandler.java:209)
> at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:104)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> {code}
> Due to this, deletion of snapshot fails:
> {code:java}
> org.apache.hadoop.hbase.snapshot.SnapshotDoesNotExistException: Snapshot
> 'FOO' doesn't exist on the filesystem
> at
> org.apache.hadoop.hbase.master.snapshot.SnapshotManager.deleteSnapshot(SnapshotManager.java:312)
> at
> org.apache.hadoop.hbase.master.MasterRpcServices.deleteSnapshot(MasterRpcServices.java:694)
> at
> org.apache.hadoop.hbase.shaded.protobuf.generated.MasterProtos$MasterService$2.callBlockingMethod(MasterProtos.java)
> at org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:413)
> at org.apache.hadoop.hbase.ipc.CallRunner.run(CallRunner.java:133)
> at org.apache.hadoop.hbase.ipc.RpcExecutor$Handler.run(RpcExecutor.java:324)
> at org.apache.hadoop.hbase.ipc.RpcExecutor$Handler.run(RpcExecutor.java:304)
> {code}
> Moreover, since the test is all about dealing with multi regions, which are
> generated by splitting, as per test logs, apparently table is not getting
> split correctly:
> {code:java}
> org.apache.hadoop.hbase.regionserver.StoreUtils(123): cannot split
> hdfs://localhost:37149/user/jenkins/test-data/b2a954c8-abb9-5528-bca3-27fcaaab735c/data/default/N000006/903675aa1a12e58bb247916dd7d492dd/0/b4bdabddce254ea5a9fa3ca432e2bca3
> because midkey is the same as first or last row
> {code}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
