[
https://issues.apache.org/jira/browse/HBASE-20414?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16447506#comment-16447506
]
Ted Yu commented on HBASE-20414:
--------------------------------
bq. does it have chance the value of (HEARTBEAT_TIMEOUT-(now-start)-10) is
negative
I don't think so. That means 1990 < now - start. From the flaky test scenario
given above, there was enough buffer for the above expression to be positive.
> TestLockProcedure#testMultipleLocks may fail on slow machine
> ------------------------------------------------------------
>
> Key: HBASE-20414
> URL: https://issues.apache.org/jira/browse/HBASE-20414
> Project: HBase
> Issue Type: Test
> Reporter: Ted Yu
> Assignee: Ted Yu
> Priority: Major
> Attachments: 20414.v1.txt
>
>
> Here was recent failure :
> https://builds.apache.org/job/HBase%20Nightly/job/branch-2.0/172/testReport/junit/org.apache.hadoop.hbase.master.locking/TestLockProcedure/health_checks___yetus_jdk8_hadoop2_checks___testMultipleLocks/
> {code}
> java.lang.AssertionError: expected:<LOCKED> but was:<UNLOCKED>
> at
> org.apache.hadoop.hbase.master.locking.TestLockProcedure.sendHeartbeatAndCheckLocked(TestLockProcedure.java:221)
> at
> org.apache.hadoop.hbase.master.locking.TestLockProcedure.testMultipleLocks(TestLockProcedure.java:311)
> {code}
> In the test output, we can see this:
> {code}
> 2018-04-13 20:19:54,230 DEBUG [Time-limited test]
> locking.TestLockProcedure(225): Proc id 22 : LOCKED.
> ...
> 2018-04-13 20:19:55,529 DEBUG [Time-limited test]
> procedure2.ProcedureExecutor(865): Stored pid=26, state=RUNNABLE;
> org.apache.hadoop.hbase.master.locking.LockProcedure
> regions=a7f9adfd047350eabb199a39564ba4db,c1e297609590b707233a2f9c8bb51fa1,
> type=EXCLUSIVE
> 2018-04-13 20:19:56,230 DEBUG [ProcExecTimeout] locking.LockProcedure(207):
> Timeout failure ProcedureEvent for pid=22, state=WAITING_TIMEOUT;
> org.apache.hadoop.hbase.master.locking.LockProcedure, namespace=namespace,
> type=EXCLUSIVE, ready=false, [pid=22, state=WAITING_TIMEOUT;
> org.apache.hadoop.hbase.master.locking.LockProcedure, namespace=namespace,
> type=EXCLUSIVE]
> {code}
> After the pid=26 log, the code does this (1 second wait):
> {code}
> // Assert tables & region locks are waiting because of namespace lock.
> Thread.sleep(HEARTBEAT_TIMEOUT / 2);
> {code}
> On a slow machine (in the case above), there was only 730 msec between the
> queueing of regionsLock2 and the WAITING_TIMEOUT state of the nsLock. The 1
> second wait was too long, leading to assertion failure.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
