[
https://issues.apache.org/jira/browse/HBASE-20589?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16479190#comment-16479190
]
stack commented on HBASE-20589:
-------------------------------
What is "...new regiog state..."? You say "...node is created with state
OFFLINE...". Which node is this? The node in master memory for hbase:meta?
Sorry for all the questions. Trying to understand what has changed. Generally
we go through the hbase:meta AssignProcedure steps to confirm that the
hbase:meta is in the place it claims to be... Even unto asking the current
server to open the hbase:meta. It will reject the request saying it already
open if carrying the region.... and then we'll know its indeed where it claims
to be.
The meta state is marked as OFFLINE on new master start and it used to be what
was in zk which would have been the last state set on last hbase:meta assign?
On the patch, I'm not mad about calling finishTransition(env, regionNode) ..
from other than the last step in the assign procedure. Whats the reason for not
setting next step as finish and calling finishTransition instead? Set next
state in starttransition to REGION_TRANSITION_FINISH ? Thanks.
> Don't need to assign meta to a new RS when standby master become active
> -----------------------------------------------------------------------
>
> Key: HBASE-20589
> URL: https://issues.apache.org/jira/browse/HBASE-20589
> Project: HBase
> Issue Type: Bug
> Reporter: Guanghao Zhang
> Assignee: Guanghao Zhang
> Priority: Major
> Attachments: HBASE-20589.master.001.patch,
> HBASE-20589.master.002.patch, HBASE-20589.master.003.patch,
> HBASE-20589.master.003.patch, HBASE-20589.master.004.patch
>
>
> I found this problem when I write ut for HBASE-20569. Now the master
> finishActiveMasterInitialization introduce a new
> RecoverMetaProcedure(HBASE-18261) and it has a sub procedure AssignProcedure.
> AssignProcedure will skip assign a region when regions state is OPEN and
> server is online. But for the new regiog state node is created with state
> OFFLINE. So it will assign the meta to a new RS. And kill the old RS when old
> RS report to master. This will make the master initialization cost a long
> time. I will attatch a ut to show this. FYI [~stack]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
