[
https://issues.apache.org/jira/browse/HBASE-20700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16506036#comment-16506036
]
Duo Zhang edited comment on HBASE-20700 at 6/8/18 1:43 PM:
-----------------------------------------------------------
Yes, the new 'meta' queue will always be served first. Typically there will be
no procedures in it so I do not think we will slow down the scheduler, just a
small number of in memory operations...
And in RecoverMetaProcedure we will do the log splitting work and assign meta,
so ideally there is no problem if the meta is on a crash RS before. Of course
the logic here is not clear enough I'd say and there maybe races, we can file
new issue to fix it.
And for the new lock, we need to prevent two RMPs run at the same time so I
think we need it...
was (Author: apache9):
Yes, the new 'meta' queue will always be served first.
And in RecoverMetaProcedure we will do the log splitting work and assign meta,
so ideally there is no problem if the meta is on a crash RS before. Of course
the logic here is not clear enough I'd say and there maybe races, we can file
new issue to fix it.
And for the new lock, we need to prevent two RMPs run at the same time so I
think we need it...
> Move meta region when server crash can cause the procedure to be stuck
> ----------------------------------------------------------------------
>
> Key: HBASE-20700
> URL: https://issues.apache.org/jira/browse/HBASE-20700
> Project: HBase
> Issue Type: Sub-task
> Components: master, proc-v2, Region Assignment
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Priority: Critical
> Fix For: 3.0.0, 2.1.0, 2.0.1
>
> Attachments: HBASE-20700-UT.patch, HBASE-20700.patch
>
>
> As said in HBASE-20682.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
