[ 
https://issues.apache.org/jira/browse/HBASE-18984?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16210573#comment-16210573
 ] 

ramkrishna.s.vasudevan commented on HBASE-18984:
------------------------------------------------

{code}
// No need to update meta here, and this is the only place we set regionNode as 
OPENING
1438          // If master crashed before #markRegionAsOpened() (where we set 
region as OPEN in meta)
1439          // And after master is restarted, this region maybe reloaded and 
assigned when master do bulk assign.
1440          // So there are might be 2 AssignProcedures for this region. One 
is created by master when bulk
1441          // assign, the other is the one failed before master crashed. But 
there can only be one
1442          // AssignProcedure per region running at a time since each 
procedure takes a lock on the region.
1443          // No matter which one AssignProcedure succeed, the the other 
will fail, and wont affect others.
{code}
I have not fully checked the flow because of this impact. But before I could 
just asking - so you have now not updating the META with OPENING state. So the 
code that was working by reading the META and checking this OPENING state has 
been removed now? I mean here in the comment you talk about master going down  
but what about the RS going down while OPENING? Yes I agree that in that case 
anyway a new node has to be assigned for the region. Will check the patch and 
the flow and will be back here.

> [AMv2] Retain assignment does not work well in AMv2
> ---------------------------------------------------
>
>                 Key: HBASE-18984
>                 URL: https://issues.apache.org/jira/browse/HBASE-18984
>             Project: HBase
>          Issue Type: Bug
>          Components: proc-v2
>    Affects Versions: 2.0.0
>            Reporter: Yi Liang
>            Assignee: Yi Liang
>             Fix For: 2.0.0
>
>         Attachments: HBASE-18984-V1-master.patch, Screen Shot 2017-10-10 at 
> 2.24.19 PM.png
>
>
> work on 8.17 Retain assignment in 
> https://docs.google.com/document/d/1eVKa7FHdeoJ1-9o8yZcOTAQbv0u0bblBlCCzVSIn69g/edit#heading=h.epjn9nege80k
> To reproduce this error, in hbase shell:
> createTable t1 --> list_regions 't1' --> disable 't1' ---> enable 't1' --> 
> list_reigons 't1' (maybe you need to try enable/disable multiple times)
> See attached images. same region assigned to different region servers



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to