incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Julien Gagnet <julien.gag...@gmail.com>
Subject Re: Caching in BigCouch
Date Tue, 20 Sep 2011 07:40:26 GMT
No. Couchdb and Bigcouch don't provide any caching. It relies on disk caching.

So even if you see the couch process taking little memory, that
doesn't mean you won't get any improvement in adding more memory to
the box. The disk cache use a sophisticated LRU algorithm and should
provide what you're after. We recently were hitted by a performance
issue and using local disk wasn't an option, we've been able to
increase the memory up to a point where most of the db fit in the disk
caching and solve our problem.

Julien

On 19 September 2011 22:15, Monica Razdan <monica@almende.org> wrote:
>
> I'm my quest to find clustering for  couchdb using hadoop I stumbled upon this blog
>
> http://mpouttuclarke.wordpress.com/2011/01/04/couchdb-and-hadoop/
>
> The author mentions here about LRU caching being used by hadoop.  I have a huge dataset
sorted on timestamp and caching on timestamp like LRU would be important for my use case.
Does bigCouch offers such mechanism under it's hood ?
>
> Monica
>
>
>
> On Sep 19, 2011, at 10:43 PM, Robert Newson <robert.newson@gmail.com> wrote:
>
> > What do you mean by 'caching' here?
> >
> > On 19 September 2011 21:36, Monica Razdan <monica@almende.org> wrote:
> >> Hello,
> >>
> >>  I wanted to compare the caching mechanism used in Hadoop - LRU with
> >> the caching in BigCouch. However, I couldn't find details about
> >> underlying caching algorithm used in Bigcouch. Could someone please
> >> direct me to the right resource?
> >>
> >> Thanks,
> >> monica
> >>

Mime
View raw message