[
https://issues.apache.org/jira/browse/HBASE-19282?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16288885#comment-16288885
]
Anastasia Braginsky commented on HBASE-19282:
---------------------------------------------
bq. Off heap test having global off heap memstore threshold as ? (40% of 16
GB?)
Approximately. 16GB on 2 Region Servers, thus 8GB on each. 8GB were distributed
as following 1GB for on-heap stuff (HBASE_HEAPSIZE=1G) and the rest for
off-heap (HBASE_OFFHEAPSIZE=7G). Then it is defined to have 40% of memory to
memstore.
<name>hbase.regionserver.global.memstore.size</name>
<value>0.4</value>
bq. In off heap tests are u reducing the xmx accordingly?
I didn't set xmx specifically, so it might be the default value. I asked
[~anoop.hbase] on HBASE-16421 and acted according to the reply there. Anyway I
want to emphasize that the results here show the difference between CCM and CAM
usage (on- and/or off-heap). We are not about to show off-heap performance.
Therefore if off-heap is not configured optimally it is the same for CCM and
for CAM.
> CellChunkMap Benchmarking and User Interface
> --------------------------------------------
>
> Key: HBASE-19282
> URL: https://issues.apache.org/jira/browse/HBASE-19282
> Project: HBase
> Issue Type: Sub-task
> Reporter: Anastasia Braginsky
> Attachments: CCM Benchmarking.pdf
>
>
> We have made some experiments how working with CellChunkMap (CCM) influences
> the performance when running on-heap and off-heap. Based on those results it
> is suggested to tie the MSLAB usage (off-heap or on-heap) with CCM index
> usage.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
