[jira] [Commented] (HBASE-17747) Support both weak and soft object pool

[Duo Zhang (JIRA)]

    [ 
https://issues.apache.org/jira/browse/HBASE-17747?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15923680#comment-15923680
 ] 

Duo Zhang commented on HBASE-17747:
-----------------------------------

This is the GC options we used for CMS
{noformat}
      -XX:+UseConcMarkSweepGC
      -verbose:gc
      -XX:+PrintGCDetails
      -XX:+PrintHeapAtGC
      -XX:+PrintGCDateStamps
      -XX:+PrintTenuringDistribution
      -XX:+PrintSafepointStatistics
      -XX:PrintSafepointStatisticsCount=1
      -XX:PrintFLSStatistics=1
      -XX:SurvivorRatio=6
      -XX:+UseCMSCompactAtFullCollection
      -XX:CMSInitiatingOccupancyFraction=75
      -XX:+UseCMSInitiatingOccupancyOnly
      -XX:+CMSParallelRemarkEnabled
      -XX:+UseNUMA
      -XX:+CMSClassUnloadingEnabled
      -XX:CMSMaxAbortablePrecleanTime=10000
      -XX:TargetSurvivorRatio=80
      -XX:+UseGCLogFileRotation
      -XX:NumberOfGCLogFiles=100
      -XX:GCLogFileSize=128m
      -XX:CMSWaitDuration=2000
      -XX:+CMSScavengeBeforeRemark
      -XX:+PrintPromotionFailure
      -XX:ConcGCThreads=16
      -XX:ParallelGCThreads=16
      -XX:PretenureSizeThreshold=2097088
      -XX:+CMSConcurrentMTEnabled
      -XX:+ExplicitGCInvokesConcurrent
      -XX:+SafepointTimeout
      -XX:MonitorBound=16384
      -XX:-UseBiasedLocking
      -XX:MaxTenuringThreshold=3
      -XX:+ParallelRefProcEnabled
{noformat}

There are lots of options could be tuned.
{quote}
So you don't really trust JVM handling GC right (smile)?
{quote}
This is your point, not me...

And still, you need to test it with G1 as the problem you described is highly 
related to the GC algorithm. I'm not saying that with G1 your code will be 
useless or have a bad performance. But we need to know it. I'm still -1 on 
committing this until you test it with G1.

Thanks.

> Support both weak and soft object pool
> --------------------------------------
>
>                 Key: HBASE-17747
>                 URL: https://issues.apache.org/jira/browse/HBASE-17747
>             Project: HBase
>          Issue Type: Improvement
>    Affects Versions: 2.0
>            Reporter: Yu Li
>            Assignee: Yu Li
>             Fix For: 2.0
>
>         Attachments: HBASE-17747.patch, HBASE-17747.v2.patch, 
> HBASE-17747.v3.patch
>
>
> During YCSB testing on embedded mode after HBASE-17744, we found that under 
> high read load GC is quite severe even with offheap L2 cache. After some 
> investigation, we found it's caused by using weak reference in 
> {{IdReadWriteLock}}. In embedded mode the read is so quick that the lock 
> might already get promoted to the old generation when the weak reference is 
> cleared, which causes dirty card table (old reference get removed and new 
> lock object set into {{referenceCache}}, see {{WeakObjectPool#get}}) thus 
> slowing YGC. In distributed mode there'll also be more lock object created 
> with weak reference than soft reference that slowing down the processing.
> So we proposed to use soft reference for this {{IdReadWriteLock}} used in 
> cache, which won't get cleared until JVM memory is not enough, and could 
> resolve the issue mentioned above. What's more, we propose to extend the 
> {{WeakObjectPool}} to be more generate to support both weak and soft 
> reference.
> Note that the GC issue only emerges under embedded mode with DirectOperator, 
> in which case all costs on the wire is removed thus produces extremely high 
> concurrency.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)