[ 
https://issues.apache.org/jira/browse/HBASE-12035?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14143601#comment-14143601
 ] 

Enis Soztutar commented on HBASE-12035:
---------------------------------------

bq. On all you say about colocated meta and master, agree, down to I no longer 
think we should have master serve meta nor any regions at all.
I was thinking about this, and we've already disabled colocation in branch-1 
and there is no benefit of colocating right now as far as I know since it is 
not forced. For 2.0, it is not clear which approach we are going to take for 
master + meta scalability (thus may not have colocation at all). Given these, 
should we undo the colocation all together? Worth discussing in dev? 

bq. But even this seems wrong asking if table is disable everytime region moves?
We have a table for namespaces (hbase:namespace) and a table for regions 
(hbase:meta), but not a table for tables, which comes in between. If we do not 
want to have another table for tables, the table state can go inside meta as 
the first row for every table. However, we should be careful about existing 
clients not creeping out for a new meta row style. 

> Client does an RPC to master everytime a region is relocated
> ------------------------------------------------------------
>
>                 Key: HBASE-12035
>                 URL: https://issues.apache.org/jira/browse/HBASE-12035
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Enis Soztutar
>            Priority: Critical
>             Fix For: 2.0.0
>
>
> HBASE-7767 moved table enabled|disabled state to be kept in hdfs instead of 
> zookeeper. isTableDisabled() which is used in 
> HConnectionImplementation.relocateRegion() now became a master RPC call 
> rather than a zookeeper client call. Since we do relocateRegion() calls 
> everytime we want to relocate a region (region moved, RS down, etc) this 
> implies that when the master is down, the some of the clients for uncached 
> regions will be affected. 
> See HBASE-7767 and HBASE-11974 for some more background. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to