[
https://issues.apache.org/jira/browse/HBASE-18366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16090355#comment-16090355
]
Umesh Agashe commented on HBASE-18366:
--------------------------------------
Hi [~yangzhe1991], thats right.
* Same as current logic, when master restarts all system table regions will be
reassigned to highest versioned RSs. Except it will be done in a new procedure
that will recover/ initialize meta and then reassign system table regions.
* When any RS crashes, only highest versioned RSs will be considered to move
system table regions on crashed RS.
Before I start working on it, I've a question: I wonder, why this logic is not
included in AssignmentManager.processAssignQueue(). So any region assignment
will first check if the region belongs to system table and consider only
highest versioned RS as targets? I can think of concerns around regions getting
re-assigned without any admin action except of course deployment of new RS.
Thanks, Umesh
> Fix flaky test
> hbase.master.procedure.TestServerCrashProcedure#testRecoveryAndDoubleExecutionOnRsWithMeta
> ---------------------------------------------------------------------------------------------------------
>
> Key: HBASE-18366
> URL: https://issues.apache.org/jira/browse/HBASE-18366
> Project: HBase
> Issue Type: Bug
> Reporter: Umesh Agashe
> Assignee: Umesh Agashe
> Priority: Blocker
> Fix For: 2.0.0
>
>
> It worked for a few days after enabling it with HBASE-18278. But started
> failing after commits:
> 6786b2b
> 68436c9
> 75d2eca
> 50bb045
> df93c13
> It works with one commit before: c5abb6c. Need to see what changed with those
> commits.
> Currently it fails with TableNotFoundException.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
