incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Paul Davis" <paul.joseph.da...@gmail.com>
Subject Re: support for massive numbers of databases on a single node
Date Tue, 11 Nov 2008 20:44:48 GMT
Troy,

When I read that I assumed it was in reference to the work Damien did
on file descriptors. IIRC, there was mailing list discussion about
someone doing huge numbers of databases. Damien just confirmed the
number was up to 500K databases.

HTH,
Paul

On Tue, Nov 11, 2008 at 3:36 PM, Troy Kruthoff <tkruthoff@blit.com> wrote:
> I was reading the Nov 2008 board report on the wiki
> (http://wiki.apache.org/couchdb/BoardReportNovember2008) and a reference to
> "support for massive numbers of databases on a single node" caught my
> attention, as we decided a few months back to hold off on using couchdb for
> our next project because our project requires "massive amounts of databases"
> and at that time, couch created a few files for each db (in
> /usr/local/var/lib/couchdb/ by default), whereas with a RDBMS we could use a
> table-space concept and not run into file-system constraints.  I've looked,
> but can't find any other mention of this topic.  Does someone know whats up?
>  In short, with the multi-key stuff in trunk and support for lots of
> databases, I'd rather be relaxing on the couch...
>
> -- troy
>
>
>

Mime
View raw message