[
https://issues.apache.org/jira/browse/HBASE-8755?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13840733#comment-13840733
]
Himanshu Vashishtha commented on HBASE-8755:
--------------------------------------------
I agree on HLogPE comment. Yes, grouping appendNoSync before calling sync is
the right way to go. The existing one is way-off the real use.
I figured that the w/o patch col in the above table contains Feng's patch too.
I am re-running the experiments at the moment with three versions: Trunk, Trunk
+ Feng's patch, Trunk + Syncer's approach. And, all versions have that HLogPe
fix on it; will report back the numbers once done.
Going by what we are seeing here, batching sync calls is definitely the right
way IMO.
I agree that Feng Honghua's patch has been tested well enough, and I really
like the radical cleanup it does. The code reads pretty clean now, though it
involves far more number of threads and synchronization stuff (which makes it
more interesting to debug too :)), I just wanted to ensure that it is safe.
The reason I mentioned this different approach is it adds lesser number of
threads (while keeping the current behaviour), and also shows improvement with
smaller number of handlers, which to me looks like a nice win over current
FSHLog. This is still in a prototype stage, and I absolutely don't want to
block Feng's superb piece of work here. It would be good to know his thoughts
on this. Thanks.
> A new write thread model for HLog to improve the overall HBase write
> throughput
> -------------------------------------------------------------------------------
>
> Key: HBASE-8755
> URL: https://issues.apache.org/jira/browse/HBASE-8755
> Project: HBase
> Issue Type: Improvement
> Components: Performance, wal
> Reporter: Feng Honghua
> Assignee: stack
> Priority: Critical
> Attachments: 8755-syncer.patch, 8755trunkV2.txt,
> HBASE-8755-0.94-V0.patch, HBASE-8755-0.94-V1.patch, HBASE-8755-0.96-v0.patch,
> HBASE-8755-trunk-V0.patch, HBASE-8755-trunk-V1.patch,
> HBASE-8755-trunk-v4.patch
>
>
> In current write model, each write handler thread (executing put()) will
> individually go through a full 'append (hlog local buffer) => HLog writer
> append (write to hdfs) => HLog writer sync (sync hdfs)' cycle for each write,
> which incurs heavy race condition on updateLock and flushLock.
> The only optimization where checking if current syncTillHere > txid in
> expectation for other thread help write/sync its own txid to hdfs and
> omitting the write/sync actually help much less than expectation.
> Three of my colleagues(Ye Hangjun / Wu Zesheng / Zhang Peng) at Xiaomi
> proposed a new write thread model for writing hdfs sequence file and the
> prototype implementation shows a 4X improvement for throughput (from 17000 to
> 70000+).
> I apply this new write thread model in HLog and the performance test in our
> test cluster shows about 3X throughput improvement (from 12150 to 31520 for 1
> RS, from 22000 to 70000 for 5 RS), the 1 RS write throughput (1K row-size)
> even beats the one of BigTable (Precolator published in 2011 says Bigtable's
> write throughput then is 31002). I can provide the detailed performance test
> results if anyone is interested.
> The change for new write thread model is as below:
> 1> All put handler threads append the edits to HLog's local pending buffer;
> (it notifies AsyncWriter thread that there is new edits in local buffer)
> 2> All put handler threads wait in HLog.syncer() function for underlying
> threads to finish the sync that contains its txid;
> 3> An single AsyncWriter thread is responsible for retrieve all the buffered
> edits in HLog's local pending buffer and write to the hdfs
> (hlog.writer.append); (it notifies AsyncFlusher thread that there is new
> writes to hdfs that needs a sync)
> 4> An single AsyncFlusher thread is responsible for issuing a sync to hdfs
> to persist the writes by AsyncWriter; (it notifies the AsyncNotifier thread
> that sync watermark increases)
> 5> An single AsyncNotifier thread is responsible for notifying all pending
> put handler threads which are waiting in the HLog.syncer() function
> 6> No LogSyncer thread any more (since there is always
> AsyncWriter/AsyncFlusher threads do the same job it does)
--
This message was sent by Atlassian JIRA
(v6.1#6144)
