hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yu Li (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-15619) Performance regression observed: Random read(get) performance of branch-1 worse than 0.98
Date Mon, 11 Apr 2016 15:26:25 GMT

    [ https://issues.apache.org/jira/browse/HBASE-15619?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15235283#comment-15235283
] 

Yu Li commented on HBASE-15619:
-------------------------------

[~stack] and all:
Please ignore case#2 and case#3, today I just found that HBASE-14061 is not committed yet
thus not included in release 1.1.4/1.2.1, so I was actually comparing ONE_SSD read for 0.98.12+
and HOT(all HDD) for 1.1.4/1.2.1. After loading data with our 1.1.2+ to generate ONE_SSD hfiles,
no more regression observed (the issue does exist in our 1.1.2+, but that's my own problem
and will resolve it in private). Sorry for the bothering.

However, for case#1 the regression does exist, so the RPCServer part may still have some problem.
Will update the JIRA title and description.

> Performance regression observed: Random read(get) performance of branch-1 worse than
0.98
> -----------------------------------------------------------------------------------------
>
>                 Key: HBASE-15619
>                 URL: https://issues.apache.org/jira/browse/HBASE-15619
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Yu Li
>            Assignee: Yu Li
>            Priority: Critical
>
> As titled, I observed the perf regression in the final stress testing before upgrading
our online cluster to 1.x. More details as follows:
> 1. HBase version in the comparison test:
>   * 0.98: based on 0.98.12 with some backports, among which HBASE-11297 is the most important
perf-related one (especially under high stress)
>   * 1.x: checked 3 releases in total
>      1) 1.1.2 with important perf fixes/improvements including HBASE-15031 and HBASE-14465
(we planed to upgrade to 1.x since Oct. last year, by when 1.1.2 was the latest stable release
in branch-1)
>      2) 1.1.4 release
>      3) 1.2.1RC1
> 2. Test environment
>     * YCSB: 0.7.0 with [YCSB-651|https://github.com/brianfrankcooper/YCSB/pull/651] applied
>     * Client: 4 physical nodes, each with 8 YCSB instance, each instance with 100 threads
>     * Server: 1 Master with 3 RS, each RS with 256 handlers and 64G heap
>     * Hardware: 64-core CPU, 256GB Mem, 10Gb Net, 1 PCIe-SSD and 11 HDD, same hardware
for client and server
> 3. Test cases
>     * -p fieldcount=1 -p fieldlength=128 -p readproportion=1
>     * case #1: read against empty table
>     * case #2: lrucache 100% hit
>     * case #3: BLOCKCACHE=>false
> 4. Test result
> * 1.1.4 and 1.2.1 have a similar perf (less than 2% deviation) as 1.1.2+, so will only
paste comparison data of 0.98.12+ and 1.1.2+
> * per-RS Throughput(ops/s)
> ||HBaseVersion||case#1||case#2||case#3||
> |0.98.12+|383562|257493|47594|
> |1.1.2+|363050|232757|35872|
> * AverageLatency(us)
> ||HBaseVersion||case#1||case#2||case#3||
> |0.98.12+|2774|4134|22371|
> |1.1.2+|2930|4572|29690|
> It seems to me each part on the read path has perf regression: RPCServer for case#1,
lrucache read for case#2, and hfile read for case#3...



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message