[
https://issues.apache.org/jira/browse/HBASE-16890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15604261#comment-15604261
]
Duo Zhang commented on HBASE-16890:
-----------------------------------
The only possible way that AsyncFSWAL does more work on checksuming is that, it
syncs more than FSHLog. The packets sent to datanode need to be aligned with
checksum boundary. So if the data is not aligned, we will copy the unaligned
data to the beginning of the buffer and its checksum will be calucated again
next time. So if AsyncFSWAL syncs more, it is more likely that the checksum of
the same piece of data will also be calculated more times.
> 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)
