TM1 Forum

Hi Guys,

I have an interesting problem happening. I have an issue with one of our cubes (Stock). Whereby if the server is restarted, when running the stock cube update process for the first time, it simply hangs. When I check the log I see the following error:

"Error: Prolog procedure line (5): Cell is not writeable"

However if instead update the rules (for example, just add a space) and then save, I can then run the process without any issue. The rules and feeders seem to work perfectly returning the correct figures.

Therefore basically each time the server is restarted, I need to open up the rules for the stock cube and then save and then I can run the updating process.

This was never a problem when we were on version 9.1, it only started happening when we recently upgarded to 9.4.

Any suggestions what I should check? Please let me know if you would like me to provide more details of the cube, process or rules?

Thanks,
Damien

- Delete the .blb file for the .rux file if you have one. It may be that the .blb and .rux are no longer in synch for some reason.

- if that doesn't help, as a workaround you might want to try having your TI process first do a ruleattach of the rules before the rest of the stuff. (Which would be tha same as you recompiling them after a server restart). Judging from your description that would work as a workaround, but ofcourse it is not the perfect solution.

Hi
Could you just confirm that you are doing a ViewZeroOut on line 5 in your TI?

I usually got this error in situations where I had data in a (range of) cell(s) on which subsequently or under condition a rule would apply.
To fix it I turned of the rules and as a once off cleared the data in those measures that are later calculated by the rule and turned the rule back on.

You mentioning changing versions, and this is a very long shot, leads me to the question whether you changed your configuration, in particular MaximumCubeLoadThreads?

Cheers

Thanks Guys for your feedback on this issue.. I have tried deleting the blb files relating to the my Stock cube and restarted the TM1 server and it was still the same problem. I then made a copy of the rules and deleted the RUX file and recreated the rules and restarted TM1 again.. Still the same problem. I then deleted the RUX file completely (meaning that there are no rules for the stock cube) and restarted TM1 and still the same problem!

I then checked and I can confirm that the ViewZeroOut function is being called on line 5 of the TI script. I have also added the MaximumCubeLoadThreads command in the TM1 config file and set it to 7 as I found in a backup that this was the original number in the old version of TM1. I restarted TM1 and still the same problem.

Gregor: When you say clear the data in the measures, not exactly sure what you mean? Do you mean deleting each of the elements in the Measure dimension and then adding them back again?

Further to my last post, I realised that in addition to changing the rules file, changing the TI process and saving that also has the same affect.. So by changing either the TI process or rules, I am then able to run the process.

I have also noticed that if I completely delete the RUX file and restart TM1 it also hangs, but a different place.. It now consistantly hangs when try to process the Warehouse dimension:

"Run:R(Warehouse)-(4)"

So it seems that it has nothing to do with rules and feeders.

Anyone have any other ideas of what may be causing this? Otherwise I may just have to live with this and touch the TI process each time the server is restarted.

Thanks for your help.

Regards,
Damien

Hi

ViewZeroOut is supposed to ignore the calculated cells as they cannot be cleared. In earlier versions there was a bug around this situation which caused the error message you mentioned.
So I don't mean deleting elements. What I was doing is clearing the data from cells which are later calculated.
An example would be: Cell A1 has a value stored. Then you write a rule which (under condition) calculates this cell A1.
What I would do is turn of this rule and clear the data in cell A1 (as it would magically have that old value again) with either TI (for a whole range) or manually for a particular cell and then turn the rule back on. After that the issue was gone for the TIs trying to ViewZeroOut on a view that included A1.

But from what you are describing (...deleting the rule with restart...same issue) I don't think this will help you.

I would set MaximumCubeLoadThreads=1 if you use conditional feeders.

Cheers

Hi Gregor,

Thankyou so much for the reply.. I setting "MaximumCubeLoadThreads=1" and then restarted the TM1 server and when running the process, it hung once again.. It s really strange.

Thanks anyway for your help.

Regards,
Damien

Is your problem that the process is hanging, as in it never stops, or is your problem that you are struggling with a zero out because of cells that are conditionally calculated ?

When I say the process is hanging, I mean that when i run the process, it basically just sits in TM1TOP with the following status:

Run:R(Stock)-(12) ChoreExecute (R)0(IX)0(W)0

If I watch Task Manager, I see the TM1 service slowly creaping up.

However it doesn't hang in the sense, that I am able to press the cancel button and it stops running.

If I simply go to the rules and enter anything such as a space and click save, I can then run the process and it finishes quickly.

Hope this answers your question.

Thanks for the help!