Hello,

Trying to optimize a cube, following the aggregates usages via performance
counters:
"Analysis Server:Last Query\DSN requested"
"Analysis Server:Last Query\DSN used"
"Analysis Server:Last Query\Query num"

This can help to see whether the requested aggregates actually exist as such.

This used to work ok. I saw the aggregates were not ok, so I deleted them,
recreated them with PartitionManager and full-refreshed the cube.

When running the MDX again the counters no longer return usefull values,
instead:
....
"11/24/2004 11:54:54.904" "0" "0" "27075"
"11/24/2004 11:54:55.904" "0" "0" "27353"
"11/24/2004 11:54:56.904" "0" "0" "27651"
....

where it should in fact be something like:
....
"11/24/2004 11:46:13.334" "1313123" "1313123" "19758"
"11/24/2004 11:46:16.333" "1313133" "1313133" "20187"
....


What do these "0" mean ?
I did restart Analysis Services and re-opened the MDX Sample app connection.
The mdx runs for about 2 mins. The performance counters are logging nothing
but "0"...
FWIW, in the access db that logs the queries I do see valid Dataset entries
like 1313132 etc. for the time slice that the perfomance monitor logged.

Thanks,
Tom
(I'm currently at the client's premises, so a fast response would be oh so
much appreciated ;-)

Okay, problem went away.
Probably I made a mistake in the cycle: new partitions -> restart service ->
start logging -> execute mdx

Restarted the service and now it is logging ok again.


"Tom VdP" wrote: