[
https://issues.apache.org/jira/browse/HBASE-16195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15370364#comment-15370364
]
Yu Li commented on HBASE-16195:
-------------------------------
This is just like some kind of self-protection, that we should not leave a
chance to let RS OOM. Just like the problem HBASE-16193 describes, the
additional reference in the Q might prevent GC on the chunks not needed any
more (w/o HBASE-16194, if we keep putting the same cell with a MSLAB pool, I
bet we will also fullGC or even OOM), so I think a limit on the Q length is
necessary.
Regarding {{memstoreFlushSize}}, it will read the customized value rather than
a hard-coded one (128MB), and yes it might be quite big, but maybe a big
upper-limit is still better than no limit? Thanks.
> Should not add chunk into chunkQueue if not using chunk pool in
> HeapMemStoreLAB
> -------------------------------------------------------------------------------
>
> Key: HBASE-16195
> URL: https://issues.apache.org/jira/browse/HBASE-16195
> Project: HBase
> Issue Type: Sub-task
> Reporter: Yu Li
> Assignee: Yu Li
> Attachments: HBASE-16195.patch, HBASE-16195_v2.patch,
> HBASE-16195_v3.patch
>
>
> Problem description and analysis please refer to HBASE-16193
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
