TM1 Forum

All - hope someone can help.

We have just upgraded Cognos Express to 10.1 as the previous version wouldn't work with Excel 2010 correctly. Since the upgrade the speed of the system has become unworkable. The main bugbear is that when I have calculated a cube view, (there is no one on the system but me) and none of the underlying data has changed, the system does a full recalculate even though the view is exactly the same. In addition to that, when I open a consolidation in the view, the system then goes and performs a full calculation - this is why it is unworkable. Every action prompts full calculation.

We didn't have this issue with the previous version - 9.4.

Is this a new feature of 10.1 and has anyone else had the same issue? Anyone suggest a fix??

Has anyone applied the latest patch? Did this have any adverse effects?

Thanks in Advance

Tim

Tim,


I have also noticed this, it seems that when IBM released the newest version they have made some fundamental changes and not documented it... completely unlike IBM I know...

Anyway I seemed to be able to bring it back into line by manually making sure that the VMM / VMT values in }CubeProperties were set and also added CalculationThresholdForStorage=5, the threshold part is set at 50 as standard and there was recently a topic on it you will be able to find in here. As to the VMM setting they have always had a standard number of bytes allocated if you leave it blank... I can't be certain but it almost seems as if they have reduced this.

Have a play about and see if you notice much of a difference, if not get on the phone to IBM... I very much doubt you will be the first one to lodge the complaint.
Also worth keeping in mind that the patch is supposed to be out this month, if it isn't already.

Good Luck.

We had this pop up at a couple of sites recently. Cured by setting ParallelInteraction=F in TM1S.CFG. Raised as a PMR with IBM who are not being receptive, though it's very easy to reproduce with a system which isn't ridiculously small. Look forward to a fix .

Thanks for the replies. We have just installed the Patch Set issued by IBM and this has resolved the issue. We made the Config file changes also just in case in Test but not in Live to see if the patch set works and it di. The Config file changes were not needed.

Our clients hit it with 'enterprise' TM1. I do hope IBM have a TM1 patch ready for children's Christmas stockings real soon now.

David Usherwood wrote:Our clients hit it with 'enterprise' TM1. I do hope IBM have a TM1 patch ready for children's Christmas stockings real soon now.

Unless big blue has changed to red and grown a white beard I wouldn't hold your breathe!!!!

David Usherwood wrote:Our clients hit it with 'enterprise' TM1. I do hope IBM have a TM1 patch ready for children's Christmas stockings real soon now.

I raised this back in July and was advised it has been fixed in 10.2 and therefore will not be issuing a patch.

That's good news.
(wonder why IBM support didn't tell us that? )

Thinking a bit more about this - TM1 10.2 isn't out yet.
Anybody know what's really happening on this?

We had the issue where 2 of our active forms are particularly slow in CE 10 interim fix pack 1 similar forms have been working fine in our old version of TM1.
It seems that this is slow for all users who didn’t make the active form/ macros. Even if they have admin permissions.

I found the Cognos Express 10.1 full fix pack 1 fixed this issue so it must have been a client side xcelerator issue.
So try downloading fix pack 1 released 3rd dec
http://www-933.ibm.com/support/fixcentr ... nction=all

Or is that what you installed David?

This issue came back after a few days and it looks like part of the issue was actually down to the RAM on the server. Although there was 32 GB of RAM, the FREE value was only 0.5 GB and as soon as this was used up, the system behaved like a dog, as my colleague put it.

The server hadn't been restated for some time and so after this was done overnight there was 20GB plus of FREE RAM and the model/application flies along.. So installing the Patch may have been a red herring after all...

Update for anyone interested - the issue is back and the server has plenty RAM and we have the Patch Installed..

I really wish we hadn't upgraded the version now....

Tim

timrrmuphy wrote:Update for anyone interested - the issue is back and the server has plenty RAM and we have the Patch Installed..

I really wish we hadn't upgraded the version now....

Tim

How often are you recycling the TM1 service? In any TM1 model, especially those with significant rule calculations, the amount of RAM usage is going to creep up over time and bog things down unless you have just a ridiculous amount of RAM. Most shops recycle their servers weekly, at a minimum, and often nightly. You may want to consider doing this of not already doing.