Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of jbellis@gmail.com designates
 74.125.82.172 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=qchcOHFpPmMue1hXYnKDWZLgKli1vCpL54pAfcv9TEXOpd1aFJESox05aGBqifPxJh
         fdfxEjq0YZPfDEBnP8vE2yZuqn9Xevln1IMU4A0Y6zm8BXsrxhi69xdlt3qqKU8eP/LV
         zYnJaVuKa6RjhRJYTuaRqVxqesjXprTL87aj0=
MIME-Version: 1.0
In-Reply-To: <F0A3E41E-4225-4308-AFD8-46B4E6DAEF02@ketralnis.com>
References: <6A9F8ED2-933E-491D-BEC0-A0F8D7210594@ketralnis.com>
	<AANLkTimjcdnQFN5XHr_C2GD6gAtgjML4NlG2FasBhPMD@mail.gmail.com>
	<F0A3E41E-4225-4308-AFD8-46B4E6DAEF02@ketralnis.com>
From: Jonathan Ellis <jbellis@gmail.com>
Date: Thu, 3 Jun 2010 19:15:49 -0700
Message-ID: <AANLkTikSnKPYC89Pl98E0KterUURUWP66dQvEIjYB1XS@mail.gmail.com>
Subject: Re: Effective cache size
To: user@cassandra.apache.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Thu, Jun 3, 2010 at 10:17 AM, David King <dking@ketralnis.com> wrote:
>>> So with the row cache, that first node (the primary replica) is the one=
 that has that row cached, yes?
>> No, it's the closest node as determined by snitch.sortByProximity.
>
> And with the default snitch, rack-unaware placement, random partitioner, =
and all nodes up, that's the primary replica, right?

No.  When all replicas have equal weight it's basically random.

>> any given node X will never know whether another node Y has a row cached=
 or not. =A0the overhead for communicating that level of detail would be to=
tally prohibitive. all caching does is speed the read, once a request is re=
ceived for data local to a given node. =A0no more, no less.
>
> Yes, that's my concern, but the details significantly affect the effectiv=
e size of the cache (in the afoorementioned case, the details place the eff=
ective size at either 6 million or 18 million, a 3x difference).
>
> So given CL=3D=3DONE reads, only the actually read node (which will be th=
e primary replica given the default placement strategy and snitch) will cac=
he the item, right? The checksum-checking doesn't cause the row to be cache=
d on the non-read nodes?

You have to read the data, before you can checksum it.  So on the
contrary, digest (checksum) vs data read has no effect on cache
behavior.

> If I read with CL=3D=3DQUORUM in an RF=3D=3D3 environment, do both read n=
odes them cache the item, or only the primary replica?

Both.  Which is what you want, otherwise your digest reads will cause
substantial unnecessary i/o on hot keys.

--=20
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of Riptano, the source for professional Cassandra support
http://riptano.com