[ 
https://issues.apache.org/jira/browse/HBASE-6466?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13972817#comment-13972817
 ] 

Liu Shaohui commented on HBASE-6466:
------------------------------------

[~zjushch]
Very curious about this patch. 

In current codebase,  the write of hlog is many small pieces write and sync, 
and the flush of memstore are batch write. So i think the throughput of one 
-thread flush  is larger than the throughput of write of hlog. That means one 
flush thread is enough for most cases. 

What's scenario for multi-thread memstore flush? High-throughput write without 
wal? 



> Enable multi-thread for memstore flush
> --------------------------------------
>
>                 Key: HBASE-6466
>                 URL: https://issues.apache.org/jira/browse/HBASE-6466
>             Project: HBase
>          Issue Type: Improvement
>          Components: regionserver
>    Affects Versions: 0.95.2
>            Reporter: chunhui shen
>            Assignee: chunhui shen
>            Priority: Critical
>             Fix For: 0.95.0
>
>         Attachments: 6466-v6.patch, 6466-v7.patch, HBASE-6466-v4.patch, 
> HBASE-6466-v4.patch, HBASE-6466-v5.patch, HBASE-6466.patch, 
> HBASE-6466v2.patch, HBASE-6466v3.1.patch, HBASE-6466v3.patch
>
>
> If the KV is large or Hlog is closed with high-pressure putting, we found 
> memstore is often above the high water mark and block the putting.
> So should we enable multi-thread for Memstore Flush?
> Some performance test data for reference,
> 1.test environment : 
> random writting;upper memstore limit 5.6GB;lower memstore limit 4.8GB;400 
> regions per regionserver;row len=50 bytes, value len=1024 bytes;5 
> regionserver, 300 ipc handler per regionserver;5 client, 50 thread handler 
> per client for writing
> 2.test results:
> one cacheFlush handler, tps: 7.8k/s per regionserver, Flush:10.1MB/s per 
> regionserver, appears many aboveGlobalMemstoreLimit blocking
> two cacheFlush handlers, tps: 10.7k/s per regionserver, Flush:12.46MB/s per 
> regionserver,
> 200 thread handler per client & two cacheFlush handlers, tps:16.1k/s per 
> regionserver, Flush:18.6MB/s per regionserver



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to