[HBASE-18930] TestRegionObserverInterface#testRecovery sometimes times out

• Type: Test
• Status: Resolved
• Priority: Major
• Resolution: Cannot Reproduce
• Assignee: Unassigned
• Reporter: Ted Yu
• Created: 03/Oct/17 23:11
• Updated: 02/Mar/18 19:19
• Resolved: 02/Mar/18 19:19

Description

During a test run, TestRegionObserverInterface#testRecovery timed out in table deletion (branch-2):

    } finally {
      util.deleteTable(tableName);
      table.close();
    }

2017-10-03 21:50:40,250 DEBUG [RpcServer.default.FPBQ.Fifo.handler=4,queue=0,port=35273] assignment.RegionTransitionProcedure(235): Received report OPENED seqId=0, pid=33,       ppid=32, state=RUNNABLE:REGION_TRANSITION_FINISH; AssignProcedure table=TestTable.testRecovery, region=e8d466bc47137a0d2d7b3119106dc88a; rit=OPENING, location=cn012.l42scl.      hortonworks.com,37573,1507067090360
...
2017-10-03 21:50:42,309 INFO  [ProcExecWrkr-9] procedure2.ProcedureExecutor(1340): Rolled back pid=35, state=ROLLEDBACK, exception=org.apache.hadoop.hbase.exceptions.            TimeoutIOException via master-delete-table:org.apache.hadoop.hbase.exceptions.TimeoutIOException: Timed out while waiting on regions TestTable.testRecovery,,1507067136028.       e8d466bc47137a0d2d7b3119106dc88a. in transition; DeleteTableProcedure table=TestTable.testRecovery exec-time=5mins, 0.253sec

It seems certain region of the table got stuck in transition, leading to the timeout of table deletion.