[ 
https://issues.apache.org/jira/browse/HBASE-19827?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16332767#comment-16332767
 ] 

stack commented on HBASE-19827:
-------------------------------

In a recent flakey report [1], there is the tell-tale line:

2018-01-19 17:12:51,915 WARN [ProcExecWrkr-15] 
procedure.RecoverMetaProcedure(174): RIT with ServerName null! pid=5, ppid=4, 
state=RUNNABLE:REGION_TRANSITION_DISPATCH; AssignProcedure table=hbase:meta, 
region=1588230740

This says that the new handleRIT code for meta added by parent issue has been 
triggered but we fail at the last stage because the vital ServerName is 
missing; it was actually cleared just before we started up the 
RecoverMetaProcedure by the ServerCrashProcedure. This patch removes that 
single line.

1. 
https://builds.apache.org/job/HBASE-Flaky-Tests-branch2.0/417/testReport/junit/org.apache.hadoop.hbase.master.assignment/TestAssignmentManager/testAssignWithRandExec/

 

> Addendum for Flakey TestAssignmentManager
> -----------------------------------------
>
>                 Key: HBASE-19827
>                 URL: https://issues.apache.org/jira/browse/HBASE-19827
>             Project: HBase
>          Issue Type: Sub-task
>          Components: flakey, test
>            Reporter: stack
>            Assignee: stack
>            Priority: Major
>             Fix For: 2.0.0-beta-2
>
>         Attachments: HBASE-19827.master.001.patch, 
> HBASE-19827.master.002.patch
>
>
> Addendum for parent. Rather than reopen parent for third time, doing subtask.
> So, in parent, we committed two fixes for failures seen in the random 
> assignment test portion of TestAssignmentManager. The second fix almost works 
> but in operation, a vital piece of info has been cleared because of the order 
> in which stuff runs.
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to