TM1 Forum

I'm a newbie and I've searched this forum and googled around but can't find an answer to the following question. Perhaps I'm querying the wrong way, so if you point me in the right direction, ie another post or website, I'd be happy to research on own.

Situation:
We have TM1 10.1 in both our prod and dev environments. I am working in dev re-writing rules to improve performance and create necessary calculations. More often than I'd like to admit (in the dev environment) I start a rule-save and realize that I didn't do something properly (the math is wrong, a feeder flag wasn't set, etc) but since syntactically correct the rule starts to save.

Question:
Is there a command in the Operations Console to do kill the ruleattach process when this happens? When I click on the ruleattach line, I can select 'kill' but it doesn't actually interrupt anything (as far as I can tell/wait). We have one cube whose rules take more than 4 hours to save, so if I make a mistake, I have to wait 4 hrs to have them finish saving incorrectly, fix and re-run.

Note: We use Excel/Perspectives to update/save rules, I gather this is the norm, but thought I'd mention it. Terminating excel via Task Manager doesn't interrupt/kill process either.

Since I've not found an answer to what seems like a routine question, perhaps not possible. If so, does anyone have suggestions for how best to stop the ruleattach so I can fix and re-save? I'm wary of stopping services or rebooting because A) it takes awhile for the tm1sd.exe to fully rebuild and B) I worry that something will be missing (irrational, perhaps), but if thats what I have to do, so be it.

Any guidance appreciated.

Thanks...

Thank you in advance...

angelnjj wrote:Is there a command in the Operations Console to do kill the ruleattach process when this happens? When I click on the ruleattach line, I can select 'kill' but it doesn't actually interrupt anything (as far as I can tell/wait).

AFAIK, the Operations Console is just layering on top of TM1 Top so you can only kill threads in that are doing are running a process or constructing a view. Most threads that are in 'Wait' can also be cancelled.

angelnjj wrote:We have one cube whose rules take more than 4 hours to save, so if I make a mistake, I have to wait 4 hrs to have them finish saving incorrectly, fix and re-run.

Usually, this is a feeder issue. Try commenting out, or somehow excluding, the feeders from your rule file. Then selectively add them back in as you are performing your development. You probably don't need to set all the feeders all the time to make sure one or two rules are behaving the way you want.

angelnjj wrote:Note: We use Excel/Perspectives to update/save rules, I gather this is the norm, but thought I'd mention it. Terminating excel via Task Manager doesn't interrupt/kill process either.

When you save the rule, the action is on the server side, not on the client. You'd have to kill the TM1 executable on the server via Task Manager instead of the Excel session.

So...sounds like I'm not able to kill a thread that is doing a ruleattach, just those that are running a process or making a view. Good to know. Advice on going through rules one at a time to see where log-jam is is great, will begin that process on dev as a side project/at night (for my own education too).

Thanks for the reply!

What I would do in your case is to reduce the amount of the data in the development so that rule saves are faster. Do you really need the complete model for development? I would argue that with some thought you should be able to create a representative subset of the model that would require less resources and time to save rule changes. Once you are happy that the rules are correct you can introduce the rest of the model back in. Another stratgey, if you are just testing the calcs in the rules, would be to take out the feeders. You don't need to feed a rule to see if the calculation is correct.

You can, if you wish, keep the data but reduce the feeders, since that's where the time is spent on startup.

There was an enhancement request in the days of yore to have a "check syntax" button as well as "save" on the rules. It doesn't help your exact situtation (especially as it doesn't exist!), never the less it would be very handy.....
err sorry for hijacking your thread...

Steve Rowe wrote:There was an enhancement request in the days of yore to have a "check syntax" button as well as "save" on the rules. It doesn't help your exact situtation (especially as it doesn't exist!), never the less it would be very handy.....

I'm sure that IBM would argue that a rule syntax checking button does exist and therefore this enhancement request is dealt with. Whether the function is retro-fitted into all rule editors is another matter ...