Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
MIME-Version: 1.0
In-Reply-To: <FE84AE7AAE9A2B4EA73512EED142BEF830BF7DDD@ATL02MB02.corp.local>
References: <FE84AE7AAE9A2B4EA73512EED142BEF830BF7DDD@ATL02MB02.corp.local>
Date: Mon, 19 Nov 2012 14:12:53 -0800
Message-ID: 
 <CACHzRHYbAexXN=y_7rcmoRp8efNYoPxNsKO+H=p6EfXi3NKiRw@mail.gmail.com>
Subject: Re: row cache re-fill very slow
From: Rob Coli <rcoli@palominodb.com>
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=f46d044630403857dc04cee06bb9

--f46d044630403857dc04cee06bb9
Content-Type: text/plain; charset=ISO-8859-1

On Mon, Nov 19, 2012 at 6:17 AM, Andras Szerdahelyi <
andras.szerdahelyi@ignitionone.com> wrote:

> How is the "saved row cache file" processed? Are the cached row keys
> simply iterated over and their respective rows read from SSTables -
> possibly creating random reads with small enough sstable files, if the keys
> were not stored in a manner optimised for a quick re-fill ? -  or is there
> a smarter algorithm ( i.e. scan through one sstable at a time, filter rows
> that should be in row cache )  at work and this operation is purely disk
> i/o bound ?
>

Nope, that's it. I am quite confident that in the version you are running,
it just assembles the row from disk, from the relevant SSTables, via the
more or less normal read path. The more fragmented your sstables, the more
random the i/o.

These two 1.2.x era JIRA relate to the row cache startup penalty :

https://issues.apache.org/jira/browse/CASSANDRA-4282 # multi-threaded row
cache loading at startup
https://issues.apache.org/jira/browse/CASSANDRA-3762 # improvements to the
AutoSavingCache (which is the base class of AutoSavingRowCache)

=Rob

-- 
=Robert Coli
AIM&GTALK - rcoli@palominodb.com
YAHOO - rcoli.palominob
SKYPE - rcoli_palominodb

--f46d044630403857dc04cee06bb9
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Mon, Nov 19, 2012 at 6:17 AM, Andras Szerdahelyi <span dir=3D"ltr">&lt;<=
a href=3D"mailto:andras.szerdahelyi@ignitionone.com" target=3D"_blank">andr=
as.szerdahelyi@ignitionone.com</a>&gt;</span> wrote:<br><blockquote class=
=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padd=
ing-left:1ex">


<div style=3D"word-wrap:break-word">How is the &quot;saved row cache file&q=
uot; processed? Are the cached row keys simply iterated over and their resp=
ective rows read from SSTables - possibly creating random reads with small =
enough sstable files, if the keys were not stored in a manner optimised for
 a quick re-fill ? - =A0or is there a smarter algorithm ( i.e. scan through=
 one sstable at a time, filter rows that should be in row cache ) =A0at wor=
k and this operation is purely disk i/o bound ?<br></div></blockquote><div>
<br></div><div>Nope, that&#39;s it. I am quite confident that in the versio=
n you are running, it just assembles the row from disk, from the relevant S=
STables, via the more or less normal read path. The more fragmented your ss=
tables, the more random the i/o.</div>
<div><br></div><div>These two 1.2.x era JIRA relate to the row cache startu=
p penalty :</div><div><br></div><div><a href=3D"https://issues.apache.org/j=
ira/browse/CASSANDRA-4282">https://issues.apache.org/jira/browse/CASSANDRA-=
4282</a> # multi-threaded row cache loading at startup</div>
<div><a href=3D"https://issues.apache.org/jira/browse/CASSANDRA-3762">https=
://issues.apache.org/jira/browse/CASSANDRA-3762</a> # improvements to the A=
utoSavingCache (which is the base class of AutoSavingRowCache)</div><div>
<br></div><div>=3DRob</div><div><br></div>-- <br><div>=3DRobert Coli<br></d=
iv><div>AIM&amp;GTALK  - <a href=3D"mailto:rcoli@palominodb.com" target=3D"=
_blank">rcoli@palominodb.com</a></div><div>YAHOO - rcoli.palominob</div><di=
v>SKYPE  - rcoli_palominodb</div>


--f46d044630403857dc04cee06bb9--