[
https://issues.apache.org/jira/browse/HBASE-18946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16216359#comment-16216359
]
stack commented on HBASE-18946:
-------------------------------
bq. But still in these cases the bulking mechanism is not a logical bulking
instead it depends on the timed wait and the size of the queue.
See in RPC where it is batching requests.
But if you want discernment regards where Regions go on a cluster, thats the
Balancer's job. It has all the sources to pull on. Can't it tell members of a
ReadReplica set? Can't it do lookup to see where the other replicas are out on
the cluster before it makes a plan for current replia? The AssignProcedure is
about assigning a single Procedure, nothing else. If we start bulking it up
with other concerns, we'll be back to the fuzzy AMv1 story.
When the Balancer is passed a List, could it look at the list first to find
replicas and group?
> Stochastic load balancer assigns replica regions to the same RS
> ---------------------------------------------------------------
>
> Key: HBASE-18946
> URL: https://issues.apache.org/jira/browse/HBASE-18946
> Project: HBase
> Issue Type: Bug
> Affects Versions: 2.0.0-alpha-3
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0-beta-1
>
> Attachments: HBASE-18946.patch, HBASE-18946.patch,
> TestRegionReplicasWithRestartScenarios.java
>
>
> Trying out region replica and its assignment I can see that some times the
> default LB Stocahstic load balancer assigns replica regions to the same RS.
> This happens when we have 3 RS checked in and we have a table with 3
> replicas. When a RS goes down then the replicas being assigned to same RS is
> acceptable but the case when we have enough RS to assign this behaviour is
> undesirable and does not solve the purpose of replicas.
> [~huaxiang] and [~enis].
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
