[Mondrian] Mondrian Cache

Julian Hyde jhyde at pentaho.com
Sun Mar 21 17:56:56 EDT 2010

Both of these are reasonable feature requests. We could have a logger &
given tracing level that prints out every request for a base cell (i.e. a
cell that comes from the database, not involving any calc members) and
whether it was a hit or a miss. And also print all cells retrieved by a
fetch from the database. (We often fetch more cells than were asked for.)
Cache size is simple: it is JVM max memory. However, it makes sense to print
the cache hit ratio for each query.
Please log feature requests for these. Or better, contribute them.
I am a bit worried that if we implemented these, people would start
obsessing about achieving a 100% hit ratio... but I guess that's a matter
for them and their therapist. :)


From: mondrian-bounces at pentaho.org [mailto:mondrian-bounces at pentaho.org] On
Behalf Of jeff.s.wright at thomsonreuters.com
Sent: Sunday, March 21, 2010 1:46 PM
To: mondrian at pentaho.org
Subject: RE: [Mondrian] Mondrian Cache

Sort of related...
Has anybody ever looked into a mechanism for reporting cell cache
statistics? The types of statistics I'm thinking about are:
* Some measure of cache size
* Some measure of cache effectiveness, such as hit ratio
--Jeff Wright


From: mondrian-bounces at pentaho.org on behalf of Bruno Gomes
Sent: Sun 3/21/2010 9:32 AM
To: mondrian at pentaho.org
Subject: [Mondrian] Mondrian Cache

i want know if its possible, given a determined MDX query, know if mondrian
uses the cache or search directly for the multidimensional cell in the DBMS.
In other words, to each mdx quey, i want to determine the multidimensional
cell's that will be used, and if that cells are already retained in mondrian
memory - mondrian cache - or if its necessary retrieve then from the data

Thanks for the help.
Best regards.

Bruno Gomes

-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 4866 bytes
Desc: not available
Url : http://lists.pentaho.org/pipermail/mondrian/attachments/20100321/1bd318df/attachment.bin 

More information about the Mondrian mailing list