[
https://issues.apache.org/jira/browse/HBASE-16417?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15728092#comment-15728092
]
Eshcar Hillel commented on HBASE-16417:
---------------------------------------
I am not suggesting to set eager as default.
We can have basic as default and if users want they can set eager on a per CF
basis. Eager can be recommended to CF that have high churn, or that have a
small working set at any point in time like our original sliding window
scenario. For example, we can recommend all applications that used to set a CF
to be IN_MEMORY to try and set it to eager.
With respect to chunks, in my experiments running with MSLABS even with no
chunk pool showed inferior results, however I didn't go much deeper with these
parameters and didn't do full exhaustive tests.
I think that you guys are much more familiar with all the details regarding L2
and off-heaping and which parameters can be played with and tuned, so it would
be a good idea that you will run these tests.
> In-Memory MemStore Policy for Flattening and Compactions
> --------------------------------------------------------
>
> Key: HBASE-16417
> URL: https://issues.apache.org/jira/browse/HBASE-16417
> Project: HBase
> Issue Type: Sub-task
> Reporter: Anastasia Braginsky
> Assignee: Eshcar Hillel
> Fix For: 2.0.0
>
> Attachments: HBASE-16417-benchmarkresults-20161101.pdf,
> HBASE-16417-benchmarkresults-20161110.pdf,
> HBASE-16417-benchmarkresults-20161123.pdf,
> HBASE-16417-benchmarkresults-20161205.pdf
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
