[Mondrian] Mondrian is slow and caches pretty less data
Sayan Hazra
sayan.mnit at gmail.com
Tue Sep 22 16:48:11 EDT 2015
Hello everyone,
I need to tune the mondrian performance parameters and make it use the
cache exhaustively and use less SQL queries. I did the following but still
could not manage to increase the speed for mondrian.
Although I've enough RAM, my mondrian is really slow and does not use the
RAM it should have been. The RAM is allocated as -Xmx40G in tomcat.
- I tried to explicitly use the parameters in mondrian.properties which
enables caching. Like the following(although these are defaults):
mondrian.expCache.enable=true
mondrian.rolap.EnableRolapCubeMemberCache=true
mondrian.rolap.star.disableCaching=false
Still the heap usage did not increase in the visualvm.
- Then I came across this property: mondrian.result.highCardChunkSize (I
used 113: there is no logic behind this. I just wanted to have a big value
as the default is 1 and made sure it is coprime with the
mondrian.result.limit) Default value: 1 Means: When reading high
cardinality dimension elements, number of elements read from database in
each step. When dealing with high cardinality dimensions, elements are
retrieved in blocks: first N elements are retrieved, when these elements
are read, next ones (from N+1 to 2N) are got. N number is this property.
Setting great values for this property increases performance but may
overload memory. Values should be prime with mondrian.result.limit(I used
5000000)
source: http://mondrian.sourceforge.net/head/configuration.html
Does it really help?
I even referred to the pentaho recommendations:
http://infocenter.pentaho.com/help/index.jsp?topic=%2Fperformance_tuning_guide%2Fconcept_mondrian_properties.html
- Any recommendations to tune the performance of Mondrian will be really
helpful apart from DB tuning(indexing and all)? I fear I am missing
something out here.
- Is there a way to cache the whole in the first mdx query for that
cube? It caches only a certain portion of the cube based on the requirement.
Please help.
Thanks and regards,
Sayan Hazra
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20150922/1f0689bc/attachment.html
More information about the Mondrian
mailing list