You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Matteo Bertozzi (JIRA)" <ji...@apache.org> on 2016/06/24 07:16:16 UTC

[jira] [Commented] (HBASE-16099) Procedures tests timing out

    [ https://issues.apache.org/jira/browse/HBASE-16099?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15347886#comment-15347886 ] 

Matteo Bertozzi commented on HBASE-16099:
-----------------------------------------

for the DeleteFamily test or the client or the master were in some "sleeping" state.
the client sent the operation to the master and the master quickly completed it. 
{noformat}
2016-06-23 18:47:22,522 INFO  [Time-limited test] client.HBaseAdmin$10(774): Started disable of testDeleteColumnFamily
2016-06-23 18:47:22,523 INFO  [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] master.HMaster(2023): Client=jenkins//67.195.81.145 disable testDeleteColumnFamily
2016-06-23 18:47:22,525 DEBUG [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] procedure2.ProcedureExecutor(669): Procedure DisableTableProcedure (table=testDeleteColumnFamily) id=26 owner=jenkins state=RUNNABLE:DISABLE_TABLE_PREPARE added to the store.
...
2016-06-23 18:47:22,614 DEBUG [ProcedureExecutor-0] procedure2.ProcedureExecutor(870): Procedure completed in 87msec: DisableTableProcedure (table=testDeleteColumnFamily) id=26 owner=jenkins state=FINISHED
{noformat}
but the client didn't ask for the state or the master was not able to process the request, 
because the request got to the master a minute later and the client quickly got the response. 
but the test at that point was already timing out
{noformat}
2016-06-23 18:48:21,508 DEBUG [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] master.MasterRpcServices(949): Checking to see if procedure is done procId=26
2016-06-23 18:48:21,509 INFO  [Time-limited test] client.HBaseAdmin$TableFuture(3454): Operation: DISABLE, Table Name: default:testDeleteColumnFamily completed
{noformat}

from a working test you can see the client asking for disable and checking the state quickly.
{noformat}
2016-06-23 18:47:14,457 INFO  [Time-limited test] client.HBaseAdmin$10(774): Started disable of testDeleteColumnFamilyTwice
2016-06-23 18:47:14,463 INFO  [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] master.HMaster(2023): Client=jenkins//67.195.81.145 disable testDeleteColumnFamilyTwice
2016-06-23 18:47:14,473 DEBUG [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] procedure2.ProcedureExecutor(669): Procedure DisableTableProcedure (table=testDeleteColumnFamilyTwice) id=7 owner=jenkins state=RUNNABLE:DISABLE_TABLE_PREPARE added to the store.
2016-06-23 18:47:14,477 DEBUG [ProcedureExecutor-0] lock.ZKInterProcessLockBase(226): Acquired a lock for /hbase/table-lock/testDeleteColumnFamilyTwice/write-master:608850000000004
2016-06-23 18:47:14,484 DEBUG [ProcedureExecutor-0] hbase.MetaTableAccessor(1355): Put{"totalColumns":1,"row":"testDeleteColumnFamilyTwice","families":{"table":[{"qualifier":"state","vlen":2,"tag":[],"timestamp":1466707634484}]}}
2016-06-23 18:47:14,490 INFO  [ProcedureExecutor-0] hbase.MetaTableAccessor(1700): Updated table testDeleteColumnFamilyTwice state to DISABLING in META
2016-06-23 18:47:14,492 DEBUG [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] master.MasterRpcServices(949): Checking to see if procedure is done procId=7
...
2016-06-23 18:47:14,559 DEBUG [ProcedureExecutor-0] procedure2.ProcedureExecutor(870): Procedure completed in 90msec: DisableTableProcedure (table=testDeleteColumnFamilyTwice) id=7 owner=jenkins state=FINISHED
2016-06-23 18:47:14,603 DEBUG [RpcServer.FifoWFPBQ.default.handler=4,queue=0,port=60885] master.MasterRpcServices(949): Checking to see if procedure is done procId=7
{noformat}

> Procedures tests timing out
> ---------------------------
>
>                 Key: HBASE-16099
>                 URL: https://issues.apache.org/jira/browse/HBASE-16099
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Appy
>
> Just triaging some tests failures upstream:
> https://builds.apache.org/view/All/job/HBase-Trunk_matrix/jdk=latest1.8,label=yahoo-not-h2/1101/testReport/junit/org.apache.hadoop.hbase.master.procedure/TestDeleteColumnFamilyProcedure/testDeleteColumnFamily/
> https://builds.apache.org/view/All/job/HBase-Trunk_matrix/jdk=latest1.8,label=yahoo-not-h2/1100/ (executor died in this one)



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)