[
https://issues.apache.org/jira/browse/HBASE-14920?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15272754#comment-15272754
]
Anoop Sam John commented on HBASE-14920:
----------------------------------------
So how will it behave when system call region flush before a graceful stop
(region close)? This will happen before split, after a replay etc? Then also
we wont do the full flush?
I agree to the point that in ur use case where mostly increment mutations,
delaying the flush of cells as much as possible helps you. Considering the
general usage where most of the cells will survive a flush, dont you think
this part only flush will make more number of small sized files and so more
#compactions? The normal decision making of when 2 flush is depending on the
memstore size. This feature helps to reduce the overall memstore heap size.
Say we will in memory flush when 25% of memstore size reaches. This will reduce
the heap size as it avoid the Cell object and CSLM entry overhead. Also this
will help to provide bigger value for memstore (provided we have more -Xmx
value).. In normal memstore the more the value for memstore, the more #cells
it will hold and when it increases CSLM perf become poor. But in this
feature, we do clear off the CSLM in btw and so we are better. So these way it
already helps much to hold off cells for longer time in memory. Am still not
convinced on the part that an explicit need/call to flush is still not flushing
the whole memstore.
> Compacting Memstore
> -------------------
>
> Key: HBASE-14920
> URL: https://issues.apache.org/jira/browse/HBASE-14920
> Project: HBase
> Issue Type: Sub-task
> Reporter: Eshcar Hillel
> Assignee: Eshcar Hillel
> Attachments: HBASE-14920-V01.patch, HBASE-14920-V02.patch,
> HBASE-14920-V03.patch, HBASE-14920-V04.patch, HBASE-14920-V05.patch,
> HBASE-14920-V06.patch, HBASE-14920-V07.patch, move.to.junit4.patch
>
>
> Implementation of a new compacting memstore with non-optimized immutable
> segment representation
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
