[
https://issues.apache.org/jira/browse/HBASE-21006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16568474#comment-16568474
]
Hari Sekhon edited comment on HBASE-21006 at 8/3/18 4:58 PM:
-------------------------------------------------------------
Data locality % improves very, very slowly over the course of days after
rolling restarts, eg. a few percent a day until it comes back up to around 90+%.
I suspect this is due to minor compactions re-writing blocks locally and not
due to region migrations, which would mean that the balancer isn't optimising
data locality back up by moving regions back where they started after a rolling
restart across the cluster.
was (Author: harisekhon):
Data locality % improves very, very slowly over the course of days after
rolling restarts, eg. a few percent a day until it comes back up to around 90+%.
I suspect this is due to minor compactions re-writing blocks locally and not
due to region migrations, which would mean that the balancer isn't optimising
data locality back up.
> Balancer - data locality drops hugely after rolling restarts on cluster, not
> factoring in data locality enough?
> ---------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-21006
> URL: https://issues.apache.org/jira/browse/HBASE-21006
> Project: HBase
> Issue Type: Improvement
> Components: Balancer
> Affects Versions: 1.1.2
> Environment: HDP 2.6
> Reporter: Hari Sekhon
> Priority: Major
>
> After doing rolling restarts of my HBase cluster the data locality drops by
> 30-40% every time which implies the stochastic balancer is not optimizing for
> data locality enough, at least not under the circumstance of rolling restarts.
> The stochastic balancer is supposed to take data locality in to account but
> if this is the case, surely it should move regions back to their original
> RegionServer and data locality should return back to around where it was, not
> drop by 30-40% percent every time I need to do some tuning and a rolling
> restarts.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
