Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (nike.apache.org: domain of scherler@gmail.com designates
 209.85.212.174 as permitted sender)
Message-ID: <52457C0C.4040405@gmail.com>
Date: Fri, 27 Sep 2013 14:37:32 +0200
From: Thorsten Scherler <scherler@gmail.com>
User-Agent: Mozilla/5.0 (X11; Linux x86_64;
 rv:17.0) Gecko/20130804 Thunderbird/17.0.8
MIME-Version: 1.0
To: user@couchdb.apache.org
Subject: Re: database limits: how many dbs can I have in one couchdb instance
References: <52455FFD.9000506@gmail.com>
 <etPan.52456197.643c9869.8980@akai.jsonified.com>
In-Reply-To: <etPan.52456197.643c9869.8980@akai.jsonified.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On 09/27/2013 12:44 PM, Dave Cottlehuber wrote:
>> Hi all,
>>
>> meanwhile the patch for "validate_doc_read" is not applied I am studying
>> different workarounds.
>>
>> One valid solution for our use case is to create a db for each user and
>> store the user specific stuff in that db.
>> That has the benefit that we can limit the access to that db with the
>> standard security mechanism and we do not
>> need to fall back to middleware logic.
>>
>> The biggest problem I see in the approach is that we need to know what
>> is the max number of "user" db
>> we can have in one instance. Regarding the size of this db it is
>> expected to be a couple of smaller docs.
>>
>> Somebody has experience with that? Like rule of thumb do not deploy more
>> then 10k databases in one instance ...
>>
>> TIA for any thoughts.
>>
>> salu2
>>
>> --
>> Thorsten Scherler  
>> codeBusters S.L. - web based systems
>>
>>
>> http://www.codebusters.es/
> Hi Thorsten,
>
> The number of DBs doesn't really matter, barring sane OS limits of files/dir.
>
> CouchDB uses an LRU cache to keep an appropriate number of DB handles
> open for active connections.
>
> A better set of questions to ask is:
>
> - how many active concurrent users (=== DBs) do you expect?
> - what sort of replication (esp filtered, continuous …) do you expect?
>
> A+
> Dave
>
>

Actually it is hard to estimate the number of user since we are building
a "online web-app generator and personal crawler". I would estimate a
exponential grow rate, but in the beginning the number of user would be
quite low and as well the growth rate is slow in the start. Regarding
the replication we would have the production system a backup server that
does continuous replication. We are still in the design phase for the
overall architecture.

salu2

-- 
Thorsten Scherler <scherler.at.gmail.com>
codeBusters S.L. - web based systems
<consulting, training and solutions>

http://www.codebusters.es/