[
https://issues.apache.org/jira/browse/HBASE-22387?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16851639#comment-16851639
]
Zheng Hu edited comment on HBASE-22387 at 5/31/19 11:35 AM:
------------------------------------------------------------
Upload three profile svg graphs with our HBASE-21879 branch
(buffer.size=65KB), Each dimension(CPU/heap allocation/lock) shows that our
code does not cause serious bottlenecks, seems the most import bottleneck is
domain-socket reading now.
The org/apache/hadoop/hbase/util/ByteBufferArray.internalTransfer is a
cpu-consuming method now, it's an async method, though won't impact the
latency or throughtput now , I will file an issue to address this.
1. cpu profile:
https://issues.apache.org/jira/secure/attachment/12970294/async-prof-pid-25042-cpu-1.svg
2. allocation profile:
https://issues.apache.org/jira/secure/attachment/12970295/async-prof-pid-25042-alloc-2.svg
3. lock profile:
https://issues.apache.org/jira/secure/attachment/12970293/async-prof-pid-25042-lock-3.svg
The graphs are provided from feature HBASE-21926, thanks
[[email protected]].
Update:
After applied the patch
[0001-Initialize-a-small-buffer-to-send-the-short-circuit-.patch|https://issues.apache.org/jira/secure/attachment/12970333/0001-Initialize-a-small-buffer-to-send-the-short-circuit-.patch],
the dfs client heap allocation dropped from 45% to 27%, the heap allocation
flame graph:
1. Before the patch:
https://issues.apache.org/jira/secure/attachment/12970295/async-prof-pid-25042-alloc-2.svg
2. After the patch:
https://issues.apache.org/jira/secure/attachment/12970475/async-prof-pid-24534-alloc-2.svg
was (Author: openinx):
Upload three profile svg graphs with our HBASE-21879 branch
(buffer.size=65KB), Each dimension(CPU/heap allocation/lock) shows that our
code does not cause serious bottlenecks, seems the most import bottleneck is
domain-socket reading now.
The org/apache/hadoop/hbase/util/ByteBufferArray.internalTransfer is a
cpu-consuming method now, it's an async method, though won't impact the
latency or throughtput now , I will file an issue to address this.
1. cpu profile:
https://issues.apache.org/jira/secure/attachment/12970294/async-prof-pid-25042-cpu-1.svg
2. allocation profile:
https://issues.apache.org/jira/secure/attachment/12970295/async-prof-pid-25042-alloc-2.svg
3. lock profile:
https://issues.apache.org/jira/secure/attachment/12970293/async-prof-pid-25042-lock-3.svg
The graphs are provided from feature HBASE-21926, thanks
[[email protected]].
> Evaluate the get/scan performance after reading HFile block into offheap
> directly
> ---------------------------------------------------------------------------------
>
> Key: HBASE-22387
> URL: https://issues.apache.org/jira/browse/HBASE-22387
> Project: HBase
> Issue Type: Sub-task
> Reporter: Zheng Hu
> Assignee: Zheng Hu
> Priority: Major
> Attachments:
> 0001-Initialize-a-small-buffer-to-send-the-short-circuit-.patch,
> G1GC-stage-stw.case01.png, G1GC-stage-stw.case02-with-buffer-size-64KB.png,
> G1GC-stw.case01.png, G1GC-stw.case02-with-buffer-size-64KB.png,
> QPS-latency.case01.png, QPS-latency.case02-with-buffer-size-64KB.png,
> async-prof-pid-24534-alloc-2.svg, async-prof-pid-25042-alloc-2.svg,
> async-prof-pid-25042-cpu-1.svg, async-prof-pid-25042-lock-3.svg,
> blockReaderRemote.png, blocksByHFile-stack-trace.png, screenshot-1.png,
> screenshot-2.png, test-cluster-configuration-details.png
>
>
> Now, all sub-tasks has been resolved now (except the HBASE-21946 because of
> the hadoop dependency problem), will provide some performance benckmarks to
> show the latency improvement.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
