Mailing-List: contact dev-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@couchdb.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Message-Id: <5D18FDB2-3529-451B-B829-E5DE7A090F61@apache.org>
From: Jan Lehnardt <jan@apache.org>
To: dev@couchdb.apache.org
In-Reply-To: <ab8e556c0905010736s717ae340r5ae405d963ae021c@mail.gmail.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Subject: Re: Reads vs Writes and contention
Date: Wed, 13 May 2009 16:25:37 +0200
References: <ab8e556c0905010736s717ae340r5ae405d963ae021c@mail.gmail.com>


On 1 May 2009, at 16:36, Kevin Jackson wrote:

> Hi,
>
> I'm interested in any experiences people may have had with couch
> reading and writing many records at the same time and what kind of
> pitfalls there may be regarding the possible contention of
> records/documents
>
> I'm intending to write some small benchmarks, but I was wondering if
> anyone had seen any problems previously.

Sorry, a late reply. In a nutshell, if you expect a lot of writes to a  
single
logical piece of data, it makes sense to split up the piece into several
documents, a new one per write and using views to fetch all related
docs for your entity in one go.

On a lower level, lightning-speed can be achieved if you keep your
harddrive head writing without seeks. For that you want a big enough
FS cache to serve your read requests and potentially an HTTP cache
in front of CouchDB for additional speedups.

Cheers
Jan
--