jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ard Schrijvers <a.schrijv...@onehippo.com>
Subject Re: search queries and caching
Date Wed, 18 May 2011 10:37:02 GMT
On Wed, May 18, 2011 at 11:27 AM, Lukas Kahwe Smith <mls@pooteeweet.org> wrote:
> Hi,
> I didnt find a good answer with some searching, my question is how search queries (SQL2,
XPath, OQM) relate to the Jackrabbit cache infrastructure.
> Am I understanding things properly that the searches themselves are not cached by Jackrabbit,
they are just send to Lucene which does its own caching? However when the search results are
then "enriched" with the actual JCR nodes, Jackrabbit will use the cache to find the nodes.
In other words Jackrabbit only has a node cache, which is used when iterating over the Lucene
search results by Jackrabbit before returning the result set to the client.

Although Jackrabbit also has some custom extra Lucene caches, your
assumptions are correct.

> Also what algorithm is used to determine what nodes should be removed from the cache
if the maximum cache size is reached?

I never looked into the caching of nodes, but I presume some LRU kind of cache

> Finally what kind of cache size do people use in production?

Someone else has to chime in here.

Regards Ard

> regards,
> Lukas Kahwe Smith
> mls@pooteeweet.org

View raw message