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 DCB2478A6 for ; Wed, 5 Oct 2011 01:25:50 +0000 (UTC) Received: (qmail 34037 invoked by uid 500); 5 Oct 2011 01:25:48 -0000 Delivered-To: apmail-cassandra-user-archive@cassandra.apache.org Received: (qmail 33993 invoked by uid 500); 5 Oct 2011 01:25:48 -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 33985 invoked by uid 99); 5 Oct 2011 01:25:48 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 05 Oct 2011 01:25:48 +0000 X-ASF-Spam-Status: No, hits=1.5 required=5.0 tests=HTML_MESSAGE,RCVD_IN_DNSWL_LOW,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: domain of joaquin@datastax.com designates 209.85.212.44 as permitted sender) Received: from [209.85.212.44] (HELO mail-vw0-f44.google.com) (209.85.212.44) by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 05 Oct 2011 01:25:42 +0000 Received: by vws5 with SMTP id 5so1163065vws.31 for ; Tue, 04 Oct 2011 18:25:21 -0700 (PDT) Received: by 10.52.75.136 with SMTP id c8mr1829634vdw.316.1317777921215; Tue, 04 Oct 2011 18:25:21 -0700 (PDT) MIME-Version: 1.0 Received: by 10.220.170.9 with HTTP; Tue, 4 Oct 2011 18:25:01 -0700 (PDT) X-Originating-IP: [64.132.24.216] In-Reply-To: References: <633818413-1317776397-cardhu_decombobulator_blackberry.rim.net-1997501838-@b4.c2.bise9.blackberry> From: Joaquin Casares Date: Tue, 4 Oct 2011 20:25:01 -0500 Message-ID: Subject: Re: EC2 raid0 disks ? To: user@cassandra.apache.org Content-Type: multipart/alternative; boundary=20cf3071ca4ae30de004ae8314d1 X-Virus-Checked: Checked by ClamAV on apache.org --20cf3071ca4ae30de004ae8314d1 Content-Type: text/plain; charset=Big5 Content-Transfer-Encoding: quoted-printable Not a problem! We ran a test a few months back and know it's better to use RAID0 vs. just one mount directory. It slips my mind whether the test was also run for multiple directories vs. RAID0 as well. We chose the RAID0 method for the AMI as to avoid confusion and allow for all the sstables to be on one drive and easier to find. This is also in lin= e with many setups that we see in the wild. Thanks, Joaquin Casares DataStax Software Engineer/Support 2011/10/4 Yang > Thanks guys. > > btw, what is the performance difference between doing a raid0 on the > multiple ephemeral drives available, and then assign it to cassandra > data directory, vs creating a mount on each of these drives, and then > specify all of these to cassandra's data directory list? > > since these drives are all virtual, would there be any benefit at all > in doing a raid0 ? > > Yang > > 2011/10/4 Joaquin Casares : > > Hello again, > > Also, EBS volumes can be attached, but the performance issues cause oth= er > > issues when running a healthy cluster. From experience running clusters > on > > EBS volumes bring their own set of unique problems and are harder to > debug. > > Here's a quick link that provides a bit more background information on > why > > it's not the best fit for Cassandra. > > http://www.mail-archive.com/user@cassandra.apache.org/msg11022.html > > Thanks, > > Joaquin Casares > > DataStax > > Software Engineer/Support > > > > > > 2011/10/4 Yi Yang > >> > >> AFAIK it's around 450G per ephemeral disk. > >> BTW randomly you can get high performance EBS drives as well. > Performance > >> are good for DB but are random in IOps. > >> ------Original Message------ > >> From: Yang > >> To: user@cassandra.apache.org > >> ReplyTo: user@cassandra.apache.org > >> Subject: EC2 raid0 disks ? > >> Sent: Oct 5, 2011 5:01 AM > >> > >> it seems that how many virtual disks you can have is fixed: > >> > >> on m2.4xlarge you have 2 disks, while on m2.2xlarge you have only 1, > >> so I can't setup a raid0 on m2.2xlarge > >> > >> am I correct? > >> > >> Thanks > >> Yang > >> > >> =B1q=A7=DA=AA=BA BlackBerry(R) =B5L=BDu=B8=CB=B8m > > > --20cf3071ca4ae30de004ae8314d1 Content-Type: text/html; charset=Big5 Content-Transfer-Encoding: quoted-printable
Not a problem!

We ran a test a few months back and= know it's better to use RAID0 vs. just one mount directory. It slips m= y mind whether the test was also run for multiple directories vs. RAID0 as = well.

We chose the RAID0 method for the AMI as to avoid confusion = and allow for all the sstables to be on one drive and easier to find. This = is also in line with many setups that we see in the wild.

Thanks,

Joaquin Casares
DataStax=
Software Engineer/Support



2011/10/4 Yang <teddyyyy123@gmail.com><= br>
Thanks guys.

btw, what is the performance difference between doing a raid0 on the
multiple ephemeral drives available, and then assign it to cassandra
data directory, vs creating a mount on each of these drives, and then
specify all of these to cassandra's data directory list?

since these drives are all virtual, would there be any benefit at all
in doing a raid0 ?

Yang

2011/10/4 Joaquin Casares <joaqu= in@datastax.com>:
> Hello again,
> Also, EBS volumes can be attached, but the performance issues cause ot= her
> issues when running a healthy cluster. From experience running cluster= s on
> EBS volumes bring their own set of unique problems and are harder to d= ebug.
> Here's a quick link that provides a bit more background informatio= n on why
> it's not the best fit for Cassandra.
> http://www.mail-archive.com/user@cassandra.apac= he.org/msg11022.html
> Thanks,
> Joaquin Casares
> DataStax
> Software Engineer/Support
>
>
> 2011/10/4 Yi Yang <i@iyyang.com= >
>>
>> AFAIK it's around 450G per ephemeral disk.
>> BTW randomly you can get high performance EBS drives as well. Perf= ormance
>> are good for DB but are random in IOps.
>> ------Original Message------
>> From: Yang
>> To: user@cassandra.ap= ache.org
>> ReplyTo: user@cassand= ra.apache.org
>> Subject: EC2 raid0 disks ?
>> Sent: Oct 5, 2011 5:01 AM
>>
>> it seems that how many virtual disks you can have is fixed:
>>
>> on m2.4xlarge you have 2 disks, while on m2.2xlarge you have only = 1,
>> so I can't setup a raid0 on m2.2xlarge
>>
>> am I correct?
>>
>> Thanks
>> Yang
>>
>> =B1q=A7=DA=AA=BA BlackBerry(R) =B5L=BDu=B8=CB=B8m
>

--20cf3071ca4ae30de004ae8314d1--