You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Duo Zhang (JIRA)" <ji...@apache.org> on 2016/11/04 13:30:58 UTC
[jira] [Updated] (HBASE-17021) Use RingBuffer to reduce the
contention in AsyncFSWAL
[ https://issues.apache.org/jira/browse/HBASE-17021?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Duo Zhang updated HBASE-17021:
------------------------------
Attachment: HBASE-17021.patch
Use ringbuffer instead of ArrayDeque in AsyncFSWAL to reduce contention.
The rolling logic is also reimplemented. Now we will only wait for the ack of all out-going appends, will not try to write out all the pending appends before rolling.
Unify AsyncFSWAL.Payload to RingBufferTruck, and also add tracing for AsyncFSWAL.append.
Fix a simple bug that we do not reset numEntries after rolling.
I set the default WAL to asyncfs in the patch. Just want to see the precommit test result to find out bugs. Will revert to filesystem when commit.
[~stack] [~ram_krish], PTAL. Thanks.
> Use RingBuffer to reduce the contention in AsyncFSWAL
> -----------------------------------------------------
>
> Key: HBASE-17021
> URL: https://issues.apache.org/jira/browse/HBASE-17021
> Project: HBase
> Issue Type: Sub-task
> Components: wal
> Affects Versions: 2.0.0
> Reporter: Duo Zhang
> Fix For: 2.0.0
>
> Attachments: HBASE-17021.patch
>
>
> The WALPE result in HBASE-16890 shows that with disruptor's RingBuffer we can get a better performance.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)