[
https://issues.apache.org/jira/browse/HBASE-16890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15604211#comment-15604211
]
stack commented on HBASE-16890:
-------------------------------
Yeah, I'd think same amount of data. We used to do bulk of checksum calculation
at append time rather than inline w/ the sync call. That might make a
difference.
Was looking at the async provider trying to relate our new concurrency model to
the old. It is not a simple mapping. We need to guard the EventLoop and
currently it is w/ the broad synchronize on a Q (DFSClient#DFSOS has all go via
synchronized Q too but less work done inside the synchronize). We should be
able to do better but would take a bit of study (You did a load of work in here
[~Apache9] -- smile).
> Analyze the performance of AsyncWAL and fix the same
> ----------------------------------------------------
>
> Key: HBASE-16890
> URL: https://issues.apache.org/jira/browse/HBASE-16890
> Project: HBase
> Issue Type: Sub-task
> Components: wal
> Affects Versions: 2.0.0
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0
>
> Attachments: contention.png, contention_defaultWAL.png
>
>
> Tests reveal that AsyncWAL under load in single node cluster performs slower
> than the Default WAL. This task is to analyze and see if we could fix it.
> See some discussions in the tail of JIRA HBASE-15536.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
