[
https://issues.apache.org/jira/browse/HBASE-8755?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13688957#comment-13688957
]
Feng Honghua commented on HBASE-8755:
-------------------------------------
[~zjushch]: We run the same tests as yours, and below are the result:
1). One YCSB client with 5/50/200 write threads respectively
2). One RS with 300 RPC handlers, 20 regions (5 data-nodes back-end HDFS
running CDH 4.1.1)
3). row-size = 150 bytes
||client-threads ||row-count ||new-model throughput ||new-model latency
||old-model throughput||old-model latency||
|5 |200000 |3191|1.551(ms) |3172 |1.561(ms)|
|50 |2000000 |23215 |2.131(ms) |7437 |6.693(ms)|
|200 |2000000 |35793 |5.450(ms) |10816 |18.312(ms)|
A). the difference is negligible when 5 threads of YCSB client, this is because
B). new-model still has 3X+ improvement compared to old-model when threads are
50/200
Can anybody else help do the tests using the same configurations as Chunhui?
Another guess is the HDFS used by chunhui has much better performance on HLog's
write/sync, which makes the new model in HBase has less impact. Just guess.
> 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: wal
> Reporter: Feng Honghua
> Attachments: HBASE-8755-0.94-V0.patch, HBASE-8755-0.94-V1.patch,
> HBASE-8755-trunk-V0.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 is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
