Return-Path: X-Original-To: apmail-cassandra-user-archive@www.apache.org Delivered-To: apmail-cassandra-user-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 3D62317CFE for ; Mon, 29 Sep 2014 18:49:50 +0000 (UTC) Received: (qmail 93175 invoked by uid 500); 29 Sep 2014 18:49:47 -0000 Delivered-To: apmail-cassandra-user-archive@cassandra.apache.org Received: (qmail 93140 invoked by uid 500); 29 Sep 2014 18:49:47 -0000 Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: user@cassandra.apache.org Delivered-To: mailing list user@cassandra.apache.org Received: (qmail 93130 invoked by uid 99); 29 Sep 2014 18:49:47 -0000 Received: from athena.apache.org (HELO athena.apache.org) (140.211.11.136) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 29 Sep 2014 18:49:47 +0000 X-ASF-Spam-Status: No, hits=2.2 required=5.0 tests=HTML_MESSAGE,RCVD_IN_DNSWL_NONE,SPF_PASS,T_REMOTE_IMAGE X-Spam-Check-By: apache.org Received-SPF: pass (athena.apache.org: domain of matmsh@yahoo.com designates 98.139.212.180 as permitted sender) Received: from [98.139.212.180] (HELO nm21.bullet.mail.bf1.yahoo.com) (98.139.212.180) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 29 Sep 2014 18:49:41 +0000 Received: from [98.139.215.141] by nm21.bullet.mail.bf1.yahoo.com with NNFMP; 29 Sep 2014 18:49:19 -0000 Received: from [98.139.212.208] by tm12.bullet.mail.bf1.yahoo.com with NNFMP; 29 Sep 2014 18:49:19 -0000 Received: from [127.0.0.1] by omp1017.mail.bf1.yahoo.com with NNFMP; 29 Sep 2014 18:49:19 -0000 X-Yahoo-Newman-Property: ymail-3 X-Yahoo-Newman-Id: 577267.67520.bm@omp1017.mail.bf1.yahoo.com Received: (qmail 96800 invoked by uid 60001); 29 Sep 2014 18:49:19 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1412016559; bh=JMGTRXyA1kd7FpP9n7CmyATrggO3V5H3G/75+1mLoIs=; h=References:Message-ID:Date:From:Reply-To:Subject:To:In-Reply-To:MIME-Version:Content-Type; b=Dw//ZRqVOKWcMZ98phG/1UczzNGnUUrpmu2IsaRmJDZd18SeyTuS1PJ9D6n8vhnPt0svH+jkBUtWv1ahUK8ESzmLB5UUDsV8/EDu17hwBuEEV3EoP76n3dAUw0WloLkrq64r0TfK4H1JUgMmuIChOXlhDXSBiGZNIxmU92eqam4= X-YMail-OSG: rD4idgcVM1l4oeEq3pGL2vLMBo5ehxq_ihRIlgMyUuAG9O4 NObqqDeqzspep6ywKt6m4zMUpyQQ3T_lGIXJsJas4AHfyGrfoWs1fL2NCYsm 5RnRPel30Dv4PrEwKgJOpXcnExWID7aqH1s4S5Z48jWJJ6c3fLrlVhwn4g.M pTCnnsuUon4qWXd5ckmhkZ99NCtE_coqn6NfY.2qEpJpvdttuV7R.cH_Pnk4 Z.mSzIaluIdy60uamJ5gDfsRYn.030OJ.i9aFJGXPpySsLcRiLrW_VgSo70a GOGY4lXtvdP50UGzXS5MUKyHFh8sdlliOsX03ceIOK18NFRAulr7J3xuQI1w cq2QbjWD6pf6qu690Ch5snm.mu8E3lNSqmVh.8ZeH5IjmKhFUhadwObrHevn Uq_5TxNsOstcbTOnoJroQ7D262MZnLxpmvMXylQYZY7Fa.jEWEIrSUUBrXdA u875GMKo5Eg_pCMWZgcxcDOyoJ8yFDws68HP1tVdtvCcXmrbfdFNR7Yh3RHy kdsZM1Q66kbtQ7DIkCFxDYLdveq.SCGKz_pmNzO_Gd4g9aOmfjq1RysA.nJ7 ylddMGEX3mtwznaR.HlwPMGFDFzUJxRddds1yEaE1EVRXkA3yisj0glT8QkM hJ0Yxffjynkz7MwqjiguUYiYI.qkOdXCF3D7xitpnx0.ewkZBFhcjJ18DeJ9 LJ_M- Received: from [62.49.31.174] by web160604.mail.bf1.yahoo.com via HTTP; Mon, 29 Sep 2014 11:49:19 PDT X-Rocket-MIMEInfo: 002.001,SSBoYXZlIHJ1biBhIHN5c2JlbmNoICBmaWxlIGlvIHRlc3Qgb24gbXkgaG9tZSBQQyBhbmQgb2ZmaWNlIFBDLiBUaGUgcmVzdWx0IGlzICBnaXZlbiBiZWxvdy4gVGhlIHJlc3VsdCBzaG93cyBteSBvZmZpY2UgUEMgKHdpdGggYSBTU0QpIGlzIGFib3V0IDMgdGltZXMgbW9yZSBwZXJmb3JtYW50IHRoYW4gbXkgaG9tZSBQQyAod2l0aCBhIHNhdGEgaGFyZCBkaXNrKS4KCkhvbWUgUEMgOgoKZ2F1c3M6fj4gc3lzYmVuY2ggLS10ZXN0PWZpbGVpbyAtLWZpbGUtdG90YWwtc2l6ZT01MEcgcHJlcGFyZQpzeXNiZW4BMAEBAQE- X-Mailer: YahooMailWebService/0.8.203.696 References: <1411767566.82796.YahooMailNeo@web160605.mail.bf1.yahoo.com> <1411809845.35013.YahooMailNeo@web160606.mail.bf1.yahoo.com> Message-ID: <1412016559.32381.YahooMailNeo@web160604.mail.bf1.yahoo.com> Date: Mon, 29 Sep 2014 11:49:19 -0700 From: Shing Hing Man Reply-To: Shing Hing Man Subject: Re: Apache Cassandra 2.1.0 : cassandra-stress performance discrepancy between SSD and SATA drive To: Kevin Burton , "user@cassandra.apache.org" In-Reply-To: <1411809845.35013.YahooMailNeo@web160606.mail.bf1.yahoo.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="-318397788-615705020-1412016559=:32381" X-Virus-Checked: Checked by ClamAV on apache.org ---318397788-615705020-1412016559=:32381 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable I have run a sysbench file io test on my home PC and office PC. The result= is given below. The result shows my office PC (with a SSD) is about 3 tim= es more performant than my home PC (with a sata hard disk).=0A=0AHome PC := =0A=0Agauss:~> sysbench --test=3Dfileio --file-total-size=3D50G prepare=0As= ysbench 0.5: multi-threaded system evaluation benchmark=0A=0A128 files, 40= 9600Kb each, 51200Mb total=0ACreating files for the test...=0AExtra file op= en flags: 0=0ACreating file test_file.0=0ACreating file test_file.1=0ACreat= ing file test_file.2=0A.........=0ACreating file test_file.125=0ACreating f= ile test_file.126=0ACreating file test_file.127=0A53687091200 bytes written= in 626.30 seconds (81.75 MB/sec).=0Amatmsh@gauss:~> sysbench --test=3Dfile= io --file-total-size=3D50G --file-test-mode=3Drndrw --init-rng=3Don --max-t= ime=3D300 --max-requests=3D0 run=0Asysbench 0.5: multi-threaded system eva= luation benchmark=0A=0ARunning the test with following options:=0ANumber of= threads: 1=0ARandom number generator seed is 0 and will be ignored=0A=0A= =0AExtra file open flags: 0=0A128 files, 400Mb each=0A50Gb total file size= =0ABlock size 16Kb=0ANumber of IO requests: 0=0ARead/Write ratio for combin= ed random IO test: 1.50=0APeriodic FSYNC enabled, calling fsync() each 100 = requests.=0ACalling fsync() at the end of test, Enabled.=0AUsing synchronou= s I/O mode=0ADoing random r/w test=0AThreads started!=0A=0AOperations perfo= rmed: 14521 reads, 9680 writes, 30976 Other =3D 55177 Total=0ARead 226.89M= b Written 151.25Mb Total transferred 378.14Mb (1.2605Mb/sec)=0A 80.67 = Requests/sec executed=0A=0AGeneral statistics:=0A total time: = 300.0030s=0A total number of events: 24201= =0A total time taken by event execution: 186.0749s=0A response time:= =0A min: 0.00ms=0A avg: = 7.69ms=0A max: = 132.43ms=0A approx. 95 percentile: 19.57ms= =0A=0AThreads fairness:=0A events (avg/stddev): 24201.0000/0.0= 0=0A execution time (avg/stddev): 186.0749/0.00=0A=0Agauss:~> =0A=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=0AOffice PC :=0Ashing@cauchy:~> sysbench --test=3Dfileio --= file-total-size=3D50G prepare=0Asysbench 0.5: multi-threaded system evalua= tion benchmark=0A=0A128 files, 409600Kb each, 51200Mb total=0ACreating file= s for the test...=0AExtra file open flags: 0=0ACreating file test_file.0=0A= Creating file test_file.1=0ACreating file test_file.2=0ACreating file test_= file.3=0A...Creating file test_file.122=0ACreating file test_file.123=0ACre= ating file test_file.124=0ACreating file test_file.125=0ACreating file test= _file.126=0ACreating file test_file.127=0A53687091200 bytes written in 175.= 55 seconds (291.66 MB/sec).=0Acauchy:~> sysbench --test=3Dfileio --file-tot= al-size=3D50G --file-test-mode=3Drndrw --init-rng=3Don --max-time=3D300 --m= ax-requests=3D0 run=0Asysbench 0.5: multi-threaded system evaluation bench= mark=0A=0ARunning the test with following options:=0ANumber of threads: 1= =0ARandom number generator seed is 0 and will be ignored=0A=0AExtra file op= en flags: 0=0A128 files, 400Mb each=0A50Gb total file size=0ABlock size 16K= b=0ANumber of IO requests: 0=0ARead/Write ratio for combined random IO test= : 1.50=0APeriodic FSYNC enabled, calling fsync() each 100 requests.=0ACalli= ng fsync() at the end of test, Enabled.=0AUsing synchronous I/O mode=0ADoin= g random r/w test=0AThreads started!=0A=0AOperations performed: 43020 read= s, 28680 writes, 91723 Other =3D 163423 Total=0ARead 672.19Mb Written 448.= 12Mb Total transferred 1.0941Gb (3.7344Mb/sec)=0A 239.00 Requests/sec ex= ecuted=0A=0AGeneral statistics:=0A total time: = 300.0007s=0A total number of events: 71700=0A total time= taken by event execution: 7.5550s=0A response time:=0A min: = 0.00ms=0A avg: = 0.11ms=0A max: 12.89ms= =0A approx. 95 percentile: 0.22ms=0A=0AThreads fairn= ess:=0A events (avg/stddev): 71700.0000/0.00=0A execution t= ime (avg/stddev): 7.5550/0.00=0A=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=0A=0A=0A=0AShing=0A=0A=0A=0AOn Saturday, 27 = September 2014, 10:24, Shing Hing Man wrote:=0A =0A=0A= =0AHi Kevin,=0A Thanks for the reply !=0AI do not know the exact brand of= SSD in my office PC. But the SSD is only 1 year old, and it is far from = full. =0A=0AOn both of office PC and home PC, I untared Apache Cassandra 2.= 1.0 and then =0A=0A run "cassandra -f " with the default config, t= hen=0A=0A run cassandra-stress =0A=0ABoth PCs have Oracle Java 1.7.= 0_40.=0A=0AI have noticed there are some parameters for SSD in cassandra.ya= ml, which I have adjusted, but with no improvement. =0A=0A=0AIt puzzles me= Cassandra on my office PC, with far better hardware, could be 100% slowe= r than my home PC. =0A=0A=0A=0AShing=0A=0A=0A=0A=0A=0A=0A=0AOn Saturday, 27= September 2014, 5:12, Kevin Burton wrote:=0A =0A=0A= =0AWhat SSD was it? There are a lot of variability in terms of SSD perform= ance.=0A=0A1. Is it a new vs old SSD? Old SSDs can become slower if they= =E2=80=99re really worn out=0A=0A2. was the office SSD near capacity holdi= ng other data?=0A=0A3. what models were they?=0A=0ASSD !=3D SSD=E2=80=A6 t= here is a massive amount of performance variability out there.=0A=0A=E2=80= =A6 also =E2=80=A6 more data is needed. JDK versions the same? cassandra = versions the same?=0A=0Awhat about the config?=0A=0A=0AOn Fri, Sep 26, 2014= at 2:39 PM, Shing Hing Man wrote:=0A=0AHi,=0A> I have = run cassandra-stress write and cassandra-stress read on my office PC an= d on my home PC. =0A>=0A>=0A>Office PC : Intel Core i7-4479, 8 virtual cor= e, 16G RAM, 500G SSD Home PC : Intel Xeon E3-1230V3, 8 virtual core, 8G RA= M, 500G SATA disk.=0A>=0A>=0A>From the cassandra-stress result (please see = below), it seems Cassandra is more than 100% performant on my home PC than= the office PC. I am expecting the other way around, as my office PC has = much better hardware. =0A>=0A>=0A>=0A>Office : Intel Core i7-4479, 9 virt= ual cores, 16G RAM, 500G SSDcauchy:~/installed/cassandra/tools/bin> ./cassa= ndra-stress write =0A>Running with 8 threadCount=0A>Results:=0A>op rate = : 11264=0A>partition rate : 11264=0A>row rate = : 11264=0A>latency mean : 0.7=0A>latency median = : 0.4=0A>latency 95th percentile : 0.9=0A>latency 99th percent= ile : 1.6=0A>latency 99.9th percentile : 5.3=0A>latency max = : 325.3=0A>Total operation time : 00:02:40=0A>=0A>=0A>=0A>=0A>cauchy:= ~/installed/cassandra/tools/bin> ./cassandra-stress read =0A>Running with 8= threadCount=0A>Results:=0A>op rate : 13702=0A>partition = rate : 13702=0A>row rate : 13702=0A>latency mea= n : 0.5=0A>latency median : 0.5=0A>latency 95th per= centile : 0.8=0A>latency 99th percentile : 1.4=0A>latency 99.9th percen= tile : 3.4=0A>latency max : 67.1=0A>Total operation time = : 00:00:30=0A>=0A>=0A>---------------------------------------------------= =0A>--------------------------------------------------=0A>=0A>Home : Intel = Xeon E3-1230V3, 8 virtual core, 8G RAM, 500G SATA disk.=0A>=0A>=0A>matmsh@= gauss:~/installed/cassandra/tools/bin> ./cassandra-stress write=0A>Running = with 8 threadCount=0A>=0A>=0A>Results:=0A>op rate : 25181= =0A>partition rate : 25181=0A>row rate : 25181= =0A>latency mean : 0.3=0A>latency median : 0.2=0A>l= atency 95th percentile : 0.3=0A>latency 99th percentile : 0.5=0A>latenc= y 99.9th percentile : 16.7=0A>latency max : 331.0=0A>Total op= eration time : 00:03:24=0A>=0A>=0A>gauss:~/installed/cassandra/tools/b= in> ./cassandra-stress read=0A>Results:=0A>op rate : 3533= 8=0A>partition rate : 35338=0A>row rate : 35338= =0A>latency mean : 0.2=0A>latency median : 0.2=0A>l= atency 95th percentile : 0.3=0A>latency 99th percentile : 0.4=0A>latenc= y 99.9th percentile : 1.1=0A>latency max : 17.7=0A>Total oper= ation time : 00:00:30=0A>=0A>=0A>=0A>=0A>Is the above result expected = ?=0A>Thanks in advance for any suggestions !=0A>=0A>=0A>Shing=0A>=0A>=0A>= =0A>=0A>=0A=0A=0A-- =0A=0AFounder/CEO Spinn3r.com=0A=0ALocation: San Franci= sco, CA=0A=0Ablog: http://burtonator.wordpress.com=0A=E2=80=A6 or check out= my Google+ profile ---318397788-615705020-1412016559=:32381 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable
I have= run a sysbench  file io test on my home PC and office PC. The result = is  given below. The result shows my office PC (with a SSD) is about 3= times more performant than my home PC (with a sata hard disk).

Home PC :
gauss:~> sysbench --test=3Dfileio --file= -total-size=3D50G prepare
sysbench 0.5:  mult= i-threaded system evaluation benchmark

128 files, 409600Kb each, 51200Mb total
Creating files for the test...
Extra file o= pen flags: 0
Creating file test_file.0
Creating file test_file.1
Creating f= ile test_file.2
.........
Creating file test_file.125
Creating file test_fi= le.126
Creating file test_file.127
53687091200 bytes written in 626.30 seconds (81.75 MB/sec).
matmsh@gauss:~> sysbench --test=3Dfileio --file-total-size=3D50G --file-test-mode=3Drndrw --init-rn= g=3Don --max-time=3D300 --max-requests=3D0 run
sys= bench 0.5:  multi-threaded system evaluation benchmark

Running the test with following options= :
Number of threads: 1
Ra= ndom number generator seed is 0 and will be ignored


Extra file open flags= : 0
128 files, 400Mb each
50Gb total file size
Block size 16Kb
Number of IO requests: 0
Read/Write r= atio for combined random IO test: 1.50
Periodic FS= YNC enabled, calling fsync() each 100 requests.
Ca= lling fsync() at the end of test, Enabled.
Using s= ynchronous I/O mode
Doing random r/w test
Threads started!

Operations performed:  14521 reads, 9680 writes, 3097= 6 Other =3D 55177 Total
Read 226.89Mb  Writte= n 151.25Mb  Total transferred 378.14Mb  (1.2605Mb/sec)
   80.67 Requests/sec executed

General statistics:
    total time:      &nb= sp;            =        300.0030s
&nb= sp;   total number of events:      =         24201
&= nbsp;   total time taken by event execution: 186.0749s
    response time:
         min:           &nbs= p;            &= nbsp;         0.00ms
         avg:  = ;            &n= bsp;            = ;       7.69ms
 = ;        max:    &nb= sp;            =             &nb= sp;  132.43ms
     &= nbsp;   approx.  95 percentile:     = ;         19.57ms

Threads fairness:
    events (avg/stddev):     =       24201.0000/0.00
&nb= sp;   execution time (avg/stddev):   186.0749/0.00

gauss:~>
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Office PC :
shin= g@cauchy:~> sysbench --test=3Dfileio --file-total-size=3D50G prepare
sysbench 0.5:  multi-threaded system evaluation benchmark

128 files, 409600Kb each, 51200Mb total
Creating files for the test...
Extra file open flags: 0
Creating file test_fil= e.0
Creating file test_file.1
Creating file test_file.2
Creating file test= _file.3
...Creating file test_file.122
Creating file test_file.123
Creating= file test_file.124
Creating file test_file.125Creating file test_file.126
Creating file test_file.127
53687091200 bytes wri= tten in 175.55 seconds (291.66 MB/sec).
cauchy:~> sysbench --test=3Dfileio --file-total-size=3D50G --file-= test-mode=3Drndrw --init-rng=3Don --max-time=3D300 --max-requests=3D0 runsysbench 0.5:  multi-threaded system evaluatio= n benchmark

Running the = test with following options:
Number of threads: 1<= br style=3D"" class=3D"">Random number generator seed is 0 and will be igno= red

Extra file open flag= s: 0
128 files, 400Mb each
50Gb total file size
Block size 16Kb
Number of IO requests: 0
Read/Write = ratio for combined random IO test: 1.50
Periodic F= SYNC enabled, calling fsync() each 100 requests.
C= alling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing rando= m r/w test
Threads started!

Operations performed:  43020 reads, 28680= writes, 91723 Other =3D 163423 Total
Read 672.19M= b  Written 448.12Mb  Total transferred 1.0941Gb  (3.7344Mb/s= ec)
  239.00 Requests/sec executed

General statistics:
    total time:     &n= bsp;            = ;        300.0007s
    total number of events:    &nb= sp;         71700
    total time taken by event execution: 7.5550s
    response time:
       &= nbsp; min:           = ;            &n= bsp;          0.00ms
         avg: = ;            &n= bsp;            = ;        0.11ms
         max:   &nb= sp;            =             &nb= sp;    12.89ms
   &n= bsp;     approx.  95 percentile:          &nb= sp;    0.22ms

Threads fairness:
    events = (avg/stddev):           7= 1700.0000/0.00
    execution time (= avg/stddev):   7.5550/0.00
=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D



Shi= ng

=
On Saturday, 27 Septembe= r 2014, 10:24, Shing Hing Man <matmsh@yahoo.com> wrote:


Hi Kevin,
   Thanks for the reply !
I do not know the exact brand of SSD in my office PC. But the SSD is  only 1 year old,  and it= is far from full.
<= /div>
On both of office PC and home PC,= I untared Apache Cassandra 2.1.0 and then
=         run = "cassandra -f " with the default config,   then
&n= bsp;       run cassandra-stress <= /span>

Both PCs  have Oracle Java 1.7.0_40.
I = have noticed there are some parameters for SSD in cassandra.yaml, which I h= ave adjusted, but with no improvement.

= It  puzzles me Cassandra on  my office PC, with far b= etter hardware,  could be 100% slower than my home PC.


Shing





<= font class=3D"" style=3D"" face=3D"Arial" size=3D"2"> On Saturday, 27 Septe= mber 2014, 5:12, Kevin Burton <burton@spinn3r.com> wrote:


What SSD was it?  The= re are a lot of variability in terms of SSD performance.

1.  Is it a new vs old SSD?  Old SSDs can become slower if= they=E2=80=99re=0A really worn out

2.  = ;was the office SSD near capacity holding other data?

3.  what models were they?
<= br class=3D"" style=3D"" clear=3D"none">
SS= D !=3D SSD=E2=80=A6 there is a massive amount of performance variability ou= t there.

=E2=80=A6 also =E2=80=A6 more data = is needed.  JDK versions the same?  cassandra versions the same?<= /div>

what about the config?

On Fri, S= ep 26, 2014 at 2:39 PM, Shing Hing Man <matmsh@yahoo.com> wrote:
Hi,
  I have run   cassandra-stress write and  cassandra-s= tress read  on my office PC and on my home PC. =0A

Office PC : Intel= Core i7-4479, 8 virtual core, 16G RAM, 500G SSD=0AHome PC : Intel = Xeon E3-1230V3, 8 virtual core, 8G RAM, 500G SATA disk.

From t= he cassandra-stress result (please see below), it seems Cassandra is more than 100% performant on my home PC than the office PC. I am expect= ing the other way around, as my office PC has much better hardware.

=0A
Office : Intel Core i7-4479, 9 virtual core= s, 16G RAM, 500G SSD
=0A= cauchy:~/installed/cassandra/tools/bin> ./cassan= dra-stress write =0A=
Running with 8 threadCount
=0A
Results:
=0A
op rate : 1126= 4
=0A
p= artition rate : 11264
=0A
row rate : 11264
=0A
latency mean = : 0.7
=0A
latency median : 0.4
=0A
= latency 95th percentile : 0.= 9
=0A
l= atency 99th percentile : 1.6
=0A
latency 99.9th percentile : 5.3=0A
latency max = : 325.3
=0A
Total operation time : 00:02:40
=0A

= =0A

=0A
cau= chy:~/installed/cassandra/tools/bin> ./cassandra-stress read=0A
Running with 8 threadCount
=0A
Results:
= =0A
op rate : 1370= 2
=0A
partition rate = : 13702
=0A
ro= w rate : 13702
=0A
latency mean : 0.5
=0A
latency median : 0.5
=0Alatency 95th percentile : 0.8<= /div>=0A
latency 99th percentile := 1.4
=0A
latency 99.9th = percentile : 3.4
=0A
lat= ency max : 67.1
=0A
Total operation time : 00:00:30
=0A
<= br class=3D"" style=3D"" clear=3D"none">
=0A
---------------------------------------------------
--------------------------------------------------
=0A
Home : Intel Xeon E3-1230V3, 8 virtual core, 8G RAM, 500G SAT= A disk.
=0A

=0A
matmsh@gauss:~/installed/= cassandra/tools/bin> ./cassandra-stress write
=0A
Running with 8 threadCount
=0A
=0A
Results:
=0A
op rate : 25181
=0A
partition rate : 25181
= =0A
row rate : 2518= 1
=0A
latency mean = : 0.3
=0A
late= ncy median : 0.2
=0A
latency 95th percentile : 0.3
=0A
latency 99th percentile : 0.5
=0A
latency max : 33= 1.0
=0A
Total operation = time : 00:03:24
=0A

=0A
gauss:~/inst= alled/cassandra/tools/bin> ./cassandra-stress read
=0A
=0A
=0A
op rate = : 35338
=0A
<= span class=3D"" style=3D"font-family:'Sans Serif';font-size:9pt;color:#0000= 00;">partition rate : 35338
=0A
row rate : 35338
=0A
latency mean : 0.2=0A
latency median : 0.= 2
=0A
latency 95th per= centile : 0.3
=0A
late= ncy 99th percentile : 0.4
=0A
latency 99.9th percentile : 1.1
=0A
latency max : 17.7
=0A
Total operation time : 00:00:30
=0A

=0A
Is the = above result expected ?
Thanks in advance for any suggesti= ons !

Shing
<= br class=3D"" style=3D"" clear=3D"none">



--
Founder/CEO Spinn3r.com
Location: San Francisco, CA
<= div class=3D"" style=3D"">blog: http://burtonator.wordpress.com
=E2=80=A6 or check out my Google+ profile
=0A




---318397788-615705020-1412016559=:32381--