Mailing-List: contact user-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hbase.apache.org
MIME-Version: 1.0
In-Reply-To: <3F729F9F8571BB49A982A6479F2BEEA451CF9EC1@SHSMSX103.ccr.corp.intel.com>
References: <CAM7-19+n7cEiY4H9iLQ3N9V0NXppOPduZwk-hhgNLEaJfiV3kA@mail.gmail.com>
 <3F729F9F8571BB49A982A6479F2BEEA451CF9EC1@SHSMSX103.ccr.corp.intel.com>
From: =?UTF-8?B?TG/Dr2MgQ2hhbmVs?= <loic.chanel@telecomnancy.net.INVALID>
Date: Fri, 18 Nov 2016 10:02:39 +0100
Message-ID: <CAEg_P-Jf7sVxd+u7dLzecgKtZWk1vdvCp12zkJ93yGiWV_woJA@mail.gmail.com>
Subject: Re: Use experience and performance data of offheap from Alibaba
 online cluster
To: user@hbase.apache.org
Cc: "dev@hbase.apache.org" <dev@hbase.apache.org>
Content-Type: multipart/alternative; boundary=001a114108ea6bdb3d05418f910e
archived-at: Fri, 18 Nov 2016 09:02:48 -0000

--001a114108ea6bdb3d05418f910e
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Nope, same here !

Lo=C3=AFc CHANEL
System Big Data engineer
MS&T - WASABI - Worldline (Villeurbanne, France)

2016-11-18 9:54 GMT+01:00 Du, Jingcheng <jingcheng.du@intel.com>:

> Thanks Yu for the sharing, great achievements.
> It seems the images cannot be displayed? Maybe just me?
>
> Regards,
> Jingcheng
>
> From: Yu Li [mailto:carp84@gmail.com]
> Sent: Friday, November 18, 2016 4:11 PM
> To: user@hbase.apache.org; dev@hbase.apache.org
> Subject: Use experience and performance data of offheap from Alibaba
> online cluster
>
> Dear all,
>
> We have backported read path offheap (HBASE-11425) to our customized
> hbase-1.1.2 (thanks @Anoop for the help/support) and run it online for mo=
re
> than a month, and would like to share our experience, for what it's worth
> (smile).
>
> Generally speaking, we gained a better and more stable
> throughput/performance with offheap, and below are some details:
>
> 1. QPS become more stable with offheap
>
> Performance w/o offheap:
>
> [cid:part1.582d4b6424f071c]
>
> Performance w/ offheap:
>
> [cid:part2.582d4b6424f071c]
>
> These data come from our online A/B test cluster (with 450 physical
> machines, and each with 256G memory + 64 core) with real world workloads,
> it shows using offheap we could gain a more stable throughput as well as
> better performance
>
> Not showing fully online data here because for online we published the
> version with both offheap and NettyRpcServer together, so no standalone
> comparison data for offheap
>
> 2. Full GC frequency and cost
>
> Average Full GC STW time reduce from 11s to 7s with offheap.
>
> 3. Young GC frequency and cost
>
> No performance degradation observed with offheap.
>
> 4. Peak throughput of one single RS
>
> On Singles Day (11/11), peak throughput of one single RS reached 100K,
> among which 90K from Get. Plus internet in/out data we could know the
> average result size of get request is ~1KB
>
> [cid:part3.582d4b6424f071c]
>
> Offheap are used on all online machines (more than 1600 nodes) instead of
> LruCache, so the above QPS is gained from offheap bucketcache, along with
> NettyRpcServer(HBASE-15756).
> Just let us know if any comments. Thanks.
>
> Best Regards,
> Yu
>
>
>
>
>
>
>

--001a114108ea6bdb3d05418f910e--