Rules do not match RUX files

[PavoGa]

TM1 version 10.1.1

Problem: Rule files called up in the editors are not matching the actual RUX files for the given cubes.

Our TM1 planning application uses a Versions dimension with five or six elements. The two that are part of this issue are Budget (where actual budgeting takes place) and Archive where numbers are stored post budget via a TI process. Rules are not to apply to the Archive slice.

The application was written before the archiving process was needed. The rules were essentially all written to apply to all versions like so (simplified for brevity):

['FICA'] = N:[Payroll Subject to FICA] * DB(TaxRateCube, 'FICARate');

Now, we want the rule to read:

['Budget', 'FICA'] = N:[Payroll Subject to FICA] * DB(TaxRateCube, 'FICARate');
##### ( we know we can use ['Archive'] = STET or something similar but for now accept we have our reasons for not doing so)#####

As we modify these rules we are getting this behavior: The cube is behaving with the first rule in place. The RUX file shows the first rule. When opening the rules file through Architect, we are getting the second rule displayed. All this follows a server restart.

Actually, on some cubes, the rules called up through Architect are substantially different from the actual RUX file stored in the data folder.

Has anyone seen this kind of behavior before? If so, what can cause it to happen through a server restart? Thanks in advance for any help!

[jim wood]

The rux file is a physical store of the last rules edited. When you save a rule it is compiled to the cube and saved to the file. If you delete the rux file the rules are still compiled to the cube and are still there. Has somebody been excel sheets to manage the rules? It sounds like the rules have some how come out of line. If open the rule in in architect and make sure the rule is saved and there fore compiled against the cube you should be fine,

Jim.

[PavoGa]

Jim,

Thank you for the response. Have not tried deleting the rules file yet to see what happens. What we have seen is hitting the space or some other innocuous change causes the file to update properly and the rules to recompile properly.

But, should TM1 reload the rules from the stored rule file on a restart?

I was just recently switched out to a persistent desktop and have had some problems like sudden disconnections, crashes, etc. In fact, right now, I'm back on a physical desktop.

It is curious.

[Alan Kirk]

Jim,

Thank you for the response. Have not tried deleting the rules file yet to see what happens. What we have seen is hitting the space or some other innocuous change causes the file to update properly and the rules to recompile properly.

But, should TM1 reload the rules from the stored rule file on a restart?

I was just recently switched out to a persistent desktop and have had some problems like sudden disconnections, crashes, etc. In fact, right now, I'm back on a physical desktop.

It is curious.

(a) Which rules editor are you using? Classic, Advanced or .xrus?
(b) What is the data directory set to in your Options? Is it the server's data directory? If not, is there another .rux file on the path that it's set to?

[Michel Zijlema]

It could be that the contents of the .RUX and the matching .blb file are out of sync. The .rux file contains the rules that are actually applied, while the .blb file contains the rules that are displayed in the rule editor.
This situation can be solved by deleting the .blb files.

Michel

[PavoGa]

Thank you for all the responses.

Alan:
First, I normally use PM in writing rules, but in converting these to specify Budget in the addressing (LH side), I am copying from the Architect editor (I guess this is what would be referred to as Classic?) to Notepad++ and then copying them back in. Have done this lots of times before with no issues.

The data directory is set to the server's data directory.

Michel:
I did not know that was possible as I have not experienced it until now. However, that makes sense. I have pretty much straightened it out now and everything seems to be in sync.

Again, thank you all!

[kaazimraza]

Thank you for all the responses.

Alan:
First, I normally use PM in writing rules, but in converting these to specify Budget in the addressing (LH side), I am copying from the Architect editor (I guess this is what would be referred to as Classic?) to Notepad++ and then copying them back in. Have done this lots of times before with no issues.

The data directory is set to the server's data directory.

Michel:
I did not know that was possible as I have not experienced it until now. However, that makes sense. I have pretty much straightened it out now and everything seems to be in sync.

Again, thank you all!

Hi PovoGa,

I have had a similar issue a while back. In my case, our rules on some element attributes cubes would disappear in Architect every time we'd deploy any object (dim/cube/process) on the server using Performance Modeller. We could see the actual rules preserved in the .RUX file but there were no rules shown in the rules editor ( classic and advanced editor ). Upon restarting the server, the disappeared rules would show up again. Our resolution was to stop using Performance Modeller and the problem disappear. Interestingly, this happened only on some element attributes cube.

The environment we have is a TM1 model built using 9.5.2 and then recently upgraded to TM1 10.1.1.

[macsir]

I simply dislike PM.

[kaazimraza]

I simply dislike PM.

Having a bit experience of working with PM in the past on model that was built entirely in TM1 10, as well as with models that were built on earlier versions and then later upgraded to 10, I prefer it for doing some specific tasks, such as deploying objects (as you can escape a reset of TM1 server), deploying applications, generating period dimension attributes, and of course changing/updating cube structures. Having said that, I have a feeling that models built in 9.5.2(versions earlier than 10), are not fully complied with PM, or, PM works in a much better way with models built on TM 10 and onwards.

@ PavoGa

Can you try doing your the stuff you are doing by keeping PM at a bay, and using Architect alone, and see how it works. Look forward to hearing from you.

[PavoGa]

The worm turned again this issue and another symptom appeared today with PRO files. See attached.

This problem has been seen in TM1 Architect so far. Because of time constraints, we did not conduct full testing to see if other editors (PM) did the same thing, but I am inclined to go along with Michel that somehow the BLB and compiled versions of the files have become disassociated. It clears up immediately when resaved (after making some minor change to trip the change flag). Note these pro files were built by another consulting firm using the wizard and are, for the most part, left alone. The missing line, as noted, is in the Generated Statements section.

I am working on a WYSE (30 years later, we're back to dumb terminals?) thin client accessing a persistent desktop through Citrix with some problems getting kicked off or lockups occurring.

My question is has anyone had this problem and determined the cause?

Problem with TM1 PRO and RUX files in QA.docx

[Wim Gielis]

[off-topic]

Note these pro files were built by another consulting firm using the wizard

Oh no!

[/off-topic]

[kaazimraza]

Note these pro files were built by another consulting firm using the wizard and are, for the most part, left alone. The missing line, as noted, is in the Generated Statements section.

Hi PavoGa

I have experienced a similar issue in the past. My scenario was as follows ;

Any process created in Architect would act crazy whenever it was opened in Performance Modeller, or vice versa. A particular instance I can recall is the renaming of variables within the process whenever we would open it from Performance Modeller.

EDIT.

I think it's worth mentioning here that the above problem I faced was on a model that was built from scratch on TM1 10

[PavoGa]

Wim,

Yes, it is somewhat off topic because it focused on PRO files, not RUX files. I purposely limited it to RUX files to begin with, but it seemed we were having the same sympton with both file types. HOWEVER, it now seems the problem with the PRO files is different because the line missing from the PRO file is a CELLPUTN on the last line of the Data tab generated statements section.

I'll be glad to start a new topic specifically for PRO file.

PavoGa