[
https://issues.apache.org/jira/browse/HBASE-2699?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12876967#action_12876967
]
ryan rawson commented on HBASE-2699:
------------------------------------
thinking about this, the background load balancer would occasionally
run, determine the optimal moves then enter them into a 'make it
happen queue'. this queue is a series of logical operations as
multiple micro-steps, eg: close region A on server X and then open
region A on server Y (after it was closed on X of course).
it would allow the load balancer to "explain" what it wants to do in
logs and then you can evaluate 2 aspects:
- is the load balancer choosing the right thing to do? (hopefully easy
to unit test)
- given these moves, are they getting executed correctly?
The latter is a little tricky, since the real world may have changed
since the balancer ran, and the code will have to check pre-conditions
and potentially abort if they are not met.
> Reimplement load balancing to be a background process and to not use
> heartbeats
> -------------------------------------------------------------------------------
>
> Key: HBASE-2699
> URL: https://issues.apache.org/jira/browse/HBASE-2699
> Project: HBase
> Issue Type: Sub-task
> Components: master, regionserver
> Reporter: Jonathan Gray
> Priority: Critical
> Fix For: 0.21.0
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
