[
https://issues.apache.org/jira/browse/HBASE-21221?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16626498#comment-16626498
]
Mingliang Liu commented on HBASE-21221:
---------------------------------------
[~yuzhih...@gmail.com], the root cause analysis makes sense to me. For the
patch, I think the {{ex}} in {{RecordingEndpoint}} should be safely published
across multiple threads, e.g. using volatile keyword.
I'm thinking if the test would be simpler by checking the controller exception
in the {{callable}} we created and passed to the {{table.coprocessorService}}.
{code:java}
CoprocessorRpcUtils.BlockingRpcCallback<MultiRowMutationProtos.MutateRowsResponse>
rpcCallback = new CoprocessorRpcUtils.BlockingRpcCallback<>();
exe.mutateRows(controller, request, rpcCallback);
+ if (controller.failedOnException()) {
+ exceptionOnMutateRows.set(true); // exceptionOnMutateRows is
AtomicBoolean declared in the test method; will assert this value.
+ }
return rpcCallback.get();
});
{code}
> Ineffective assertion in TestFromClientSide3#testMultiRowMutations
> ------------------------------------------------------------------
>
> Key: HBASE-21221
> URL: https://issues.apache.org/jira/browse/HBASE-21221
> Project: HBase
> Issue Type: Test
> Reporter: Ted Yu
> Assignee: Ted Yu
> Priority: Minor
> Attachments: 21221.v7.txt, 21221.v8.txt, 21221.v9.txt
>
>
> Observed the following in
> org.apache.hadoop.hbase.util.TestFromClientSide3WoUnsafe-output.txt :
> {code}
> Caused by:
> org.apache.hadoop.hbase.ipc.RemoteWithExtrasException(java.io.IOException):
> java.io.IOException: Timed out waiting for lock for row: ROW-1 in region
> 089bdfa75f44d88e596479038a6da18b
> at
> org.apache.hadoop.hbase.regionserver.HRegion.getRowLockInternal(HRegion.java:5816)
> at
> org.apache.hadoop.hbase.regionserver.HRegion$4.lockRowsAndBuildMiniBatch(HRegion.java:7432)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.doMiniBatchMutate(HRegion.java:4008)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.batchMutate(HRegion.java:3982)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.mutateRowsWithLocks(HRegion.java:7424)
> at
> org.apache.hadoop.hbase.coprocessor.MultiRowMutationEndpoint.mutateRows(MultiRowMutationEndpoint.java:116)
> at
> org.apache.hadoop.hbase.protobuf.generated.MultiRowMutationProtos$MultiRowMutationService.callMethod(MultiRowMutationProtos.java:2266)
> at
> org.apache.hadoop.hbase.regionserver.HRegion.execService(HRegion.java:8182)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.execServiceOnRegion(RSRpcServices.java:2481)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.execService(RSRpcServices.java:2463)
> ...
> Exception in thread "pool-678-thread-1" java.lang.AssertionError: This cp
> should fail because the target lock is blocked by previous put
> at org.junit.Assert.fail(Assert.java:88)
> at
> org.apache.hadoop.hbase.client.TestFromClientSide3.lambda$testMultiRowMutations$7(TestFromClientSide3.java:861)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> {code}
> Here is related code:
> {code}
> cpService.execute(() -> {
> ...
> if (!threw) {
> // Can't call fail() earlier because the catch would eat it.
> fail("This cp should fail because the target lock is blocked by
> previous put");
> }
> {code}
> Since the fail() call is executed by the cpService, the assertion had no
> bearing on the outcome of the test.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
