TM1 9.1.3 on Win 2003 64 bit - MEMORY_TEMP_POOL_EXCEEDED

[att]

We started getting a strange error:

1456 WARN 2009-01-15 18:03:37,698 TM1.Server.Memory al_Alloc() outOfMemory Exception <<< MEMORY_TEMP_POOL_EXCEEDED >>> - threadID "1456" - apifunc# "134" - pool# "0" - poolsize "256.000000"
...
1456 WARN 2009-01-15 17:49:08,013 TM1.Server.Memory al_Alloc() outOfMemory Exception <<< MEMORY_TEMP_POOL_EXCEEDED >>> - threadID "1456" - apifunc# "134" - pool# "0" - poolsize "33024.000000"

We are running 64 bit TM1 9.1.3 on Windows 2003 64 bit - TM1 never used more than 7GB of RAM and the server has 20GB RAM. The error seems linked to the new rule we created - the rule itself returns correct results - we double checked it, but after the rule has been created some views stopped working producing the error above, other views still work fine even as they return more or less the same data, and some views work or crash intermittently - it seems it depends on order in which you open them.

Thank you for you help.

[George Regateiro]

When I get a chance tonight I will search through my old cases that have been Iogged. I know we received this at one point, but I cannot remember how we resolved the issue. I can remember that it was something very quirky on or end.

[Steve Rowe]

Do you fancy posting the rule and it's feeder?
It might jog George's memory.
My bet would be that there's a stack getting blown on the rule or feeder side.
Cheers,

[Armand Mizan]

We have been receiving a similar message, however in our models, hardly any rules are used.Our server however is a beast...64 bit with 128gb of ram, so seeing this error came as a bit of a surprise.

Cheers

Armand

[Steve Rowe]

@ att,
It's possible you have two issues that are not related.

You are seeing an internal warning message from the server that is not important and nothing to worry about.

and

You have a problem with your new rule and feeder
?

[George Regateiro]

Sorry for the delay, I have been working far to many hours lately. Our problem was a rule problem, but according to Cognos it had nothing to do with the error (at least directly)

From their support

Another user connecting to the server and opening a large view would cause the server to throw the error that you saw. It may make sense for us to install TM1Top if the problem persists.
Let me know how you'd like to proceed.


Since then we have tweaked our rules and dimension structures to be more efficient on the user end and have not seen this since. In our case it turned out that the rule was calculating more cells then it needed to so when the users brought back their standard views it was exceeding the max memory on the view. If it is happening very frequenlty it may be a sign of some tweaks needed to the rules.

[att]

@ att,
It's possible you have two issues that are not related.

You are seeing an internal warning message from the server that is not important and nothing to worry about.

and

You have a problem with your new rule and feeder
?

Hi All,

The problem was soved by searching this forum and including the following line into our .CFG file:

MaximumViewSize=2048

Everything works fine now.