jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Patricio Echagüe <patric...@gmail.com>
Subject Help to improve this SQL2 query
Date Wed, 10 Mar 2010 00:41:23 GMT
Hi all, I'm running this query in my repository with 5k users (nodes) where
everyone has one Application node and each application has 4 Folder nodes.

root
   |
(I have here some system folder to keep only 2k users )
   |
user 1      .......... user 5000
    \
  application
        \
   ---------------
   \     \     \     \
  F1  F2  F3   F4

For some reason the query (1) executes really fast vs the second one that
executes 20 times slower. I'm worried because the repository is almost empty
and this query shold be running as fast as (1).

(1) executed in 0.06 s. (SELECT * FROM [scrat:app] AS nodeType WHERE
nodeType.vparentid = 'd93c442a-2096-444a-9b6c-e52ffd4d88b5' AND
nodeType.[jcr:name] = 'Email')

(2) executed in 1.14 s. (SELECT * FROM [myapp:folder] AS nodeType WHERE
nodeType.vparentid =  'da392a3b-949c-4542-a75b-283664547878' AND
nodeType.[jcr:name] = 'Inbox')

My caches are by default. BundleCache in PM is 128 MB (seems ok).
I'm running against Cassandra which is extremely fast.

Am I doing something wrong? May someone suggest any way to speed it up?

thanks in advance

-- 
Patricio.-

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message