[
https://issues.apache.org/jira/browse/HBASE-20188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16404407#comment-16404407
]
Anoop Sam John commented on HBASE-20188:
----------------------------------------
bq.If having CMS makes the actual flush to cost more because of compares but at
the same time since CMS creates smaller size CSLM writes are faster, may be
should have the right balance there?
I dont think here the issue is perf numbers. The issue Stack mentioned is he
is getting so many RegionTooBusyExceptions at client side. These are thrown
from Region when the memstore size is 4x larger than flush size. Both these ,
more speed in Map write and slowness in flush will account to this. So the
impact is double( not balancing). Am I making my thinking clear now?
> [TESTING] Performance
> ---------------------
>
> Key: HBASE-20188
> URL: https://issues.apache.org/jira/browse/HBASE-20188
> Project: HBase
> Issue Type: Umbrella
> Components: Performance
> Reporter: stack
> Priority: Critical
> Fix For: 2.0.0
>
> Attachments: flamegraph-1072.1.svg, flamegraph-1072.2.svg, tree.txt
>
>
> How does 2.0.0 compare to old versions? Is it faster, slower? There is rumor
> that it is much slower, that the problem is the asyncwal writing. Does
> in-memory compaction slow us down or speed us up? What happens when you
> enable offheaping?
> Keep notes here in this umbrella issue. Need to be able to say something
> about perf when 2.0.0 ships.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
