Mailing-List: contact cassandra-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cassandra-user@incubator.apache.org
Received-SPF: pass (nike.apache.org: domain of jbellis@gmail.com designates
 74.125.82.47 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:from:date:message-id:subject:to
         :content-type:content-transfer-encoding;
        b=tcf1Kq9ld1xF8zfj6E7I1JoaCeBxc2Uk3TAlUkH2LR1GeULEx4BLZ4HAYYin3mOGm4
         FEfua/PPeBm/zH4cPxNIVS/CZiLiKfYbvI3rjkNwUWudgYrOO0o7n3hmiDIfhmZp/Xib
         /CAjshZ9rk2bpf4yHMOjIr1eehpeEtO60JNXE=
MIME-Version: 1.0
In-Reply-To: <20100225215403.GB46232@alumni.caltech.edu>
References: <20100225192342.GA46024@alumni.caltech.edu>
 <1c6fcff1002251136j21525c26hc7062d4227c4ae88@mail.gmail.com>
	<e06563881002251143k3788aec0l58affd79736402f8@mail.gmail.com>
	<20100225204832.GA46232@alumni.caltech.edu>
 <e06563881002251307o134a04b4nff706a66062a8211@mail.gmail.com>
	<20100225215403.GB46232@alumni.caltech.edu>
From: Jonathan Ellis <jbellis@gmail.com>
Date: Thu, 25 Feb 2010 22:42:54 -0600
Message-ID: <e06563881002252042ua571ebai3c64e89a48e129a3@mail.gmail.com>
Subject: Re: Multiple Data Directories
To: cassandra-user@incubator.apache.org, gdusbabek@gmail.com
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Thu, Feb 25, 2010 at 3:54 PM, Anthony Molinaro
<anthonym@alumni.caltech.edu> wrote:
> What about the case where cpu and ram are underutilized, and your bottlen=
eck
> is disk io (which seems to often be the case in ec2), then adding more
> spindles improves overall throughput of the system. =A0I've actually test=
ed
> this when adding an additional ebs, and hand moving files around, then
> restarting. =A0Suddenly node's performance (measued via cfstats metrics),
> get better

That sounds like you're actually ram-limited, so adding nodes will be
better than adding EBS devices.

>=A0How do
> the files ever get that big, does a repair fully compact (ie, down
> to one file)? =A0I guess the question is how do you end up with the
> "worst" case?

Any major compaction will do that.  Repair will invoke one, or it can
happen "naturally" too.

> I guess Raid0 is the only way to use multiple disks efficiently and
> the multiple DataFileDirectories is really not very useful? =A0I'm
> trying to think of a good reason you might want multiple data directories
> and all I can't think of one now, is there a good reason?

If you are throughput-bound instead of size-bound (which is the case
for most uses, especially on non-virtual hardware), then I would
expect better performance from JBOD.

-Jonathan