[ 
https://issues.apache.org/jira/browse/HBASE-7735?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13607173#comment-13607173
 ] 

terry zhang commented on HBASE-7735:
------------------------------------

Hi Jonathan Hsieh,  Can we use below methods to increase the probability of 
snapshots succeeding?

1. get all the region list in master. Including moving region , online region 
and spliting region(parent and child)
2. check the region list to make sure there is no hole in it.
3. generate a snapshot task assign map file in the target snapshot folder. 
(also assign the moving regions and spliting parent regions to some region 
server )
4. when the regionserver start buildSubprocedure. it will compare online region 
and the assgin region list in task file which master generated . if it is a 
online region it will be a FlushSnapshotSubprocedure. if it is not online we 
can think it is a close region. we do not need to flush cache , only need to 
create reference file(empty file). If the region already in the snapshot 
folder. we can just skip .

I think this would be helpful cause in large cluster moving region and spliting 
region is a normal stituation. So snapshot maybe always fail when do the 
verification. what do you think ?
                
> Prevent regions from moving during online snapshot.
> ---------------------------------------------------
>
>                 Key: HBASE-7735
>                 URL: https://issues.apache.org/jira/browse/HBASE-7735
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Jonathan Hsieh
>
> To increase the probability of snapshots succeeding, we should attempt to 
> prevent splits and region moves from happening.  Currently we take region 
> locks but this could be "too late" and results in an aborted snapshot.  
> We should probably take the table lock (0.96) when starting a snapshot and 
> for  a 0.94 backport we should probably disable the balancer.
> This will probably not be tackled until after trunk merge.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to