Server crash after Export to Textdata

[holger_b]

Dear forum,

before I take the effort to set up a PMR, has anyone ever encountered anything similar? We twice experienced a server crash executing the following steps:

Create a view in a TI process (process is ended by ProcessQuit immediately after this)
Export the view to a text file using Export to Textdata

The export file contains just one line, I would expect a few more. These are the last lines in the log:

10844 [453] ERROR 2012-06-04 15:51:32.415 TM1.Process Process "6014_Subversion_to_Data_copy_All_Sources_physical_copy": : Execution was aborted by ProcessQuit Function.
10844 [453] ERROR 2012-06-04 15:53:23.348 TM1.Cube db_NXTREC::initializeReferences: Failure to reference the target cube with index: '0'. Ending iteration.

This is 9.5.2, Fixpack 2.

Any hints?

Thank you
Holger

[Duncan P]

You did a ProcessQuit.

10844 [453] ERROR 2012-06-04 15:51:32.415 TM1.Process Process "6014_Subversion_to_Data_copy_All_Sources_physical_copy": : Execution was aborted by ProcessQuit Function.

Everything you did in the process should get rolled back at this point.

Why not do a ProcessBreak?

[holger_b]

Duncan, I can hardly believe that, since I do that regularly to check the outcome of processes, like the views I create. I do not use ProcessBreak since that executes the epilog tab, which I do not want.

[Delbrooks]

Hi

Check the source view. All the dimensions in the view need to be in rows column. When the source is corrupted generally then you get such an error.

Thanks
~D

[AmbPin]

Hello,

I too have seen this error and have raised it through our support vendor. IBM have been able to recreate the problem too. We only started getting this when we installed FP2.

[holger_b]

Thank you AmpBin! Can you indicate the SR number please?

Regards
Holger

[AmbPin]

I am afraid it disapears into a black hole so I don't get to see any numbers.

[kangkc]

Just got hit and whole server just went down. 9.5.2 FP2.
Exact same message in log and was doing a export text file from view.

I found the APAR opened in IBM PM68926 on the 16 Jul 2012.

My worry now is this is such a common use feature and I'll seeing TM1 crashing every other day. Goodness.

[Gabor]

I had the same issue running with 9.5.2 FP2 (HF1 ...HF3 applied). It even appeared while simply using the Textoutput to .cma in Server Explorer.
Have applied the newer Hotfixes now (HF4...HF6 / on the server only) and it seems to work fine.

[lotsaram]

I had the same issue running with 9.5.2 FP2 (HF1 ...HF3 applied). It even appeared while simply using the Textoutput to .cma in Server Explorer.
Have applied the newer Hotfixes now (HF4...HF6 / on the server only) and it seems to work fine.

I also had a PMR raised for this but kind of let it slide as a "reproducible test case" would have involved purposefully crashing a server to get a dump file. No thank you IBM. Good to know that the latest HF for FP2 seems to have this fixed, someone else must have had more perseverance with IBM support than I did, they certainly believe in the concept of "innocent until proven guilty" when it comes to recognizing bugs.

[Helmar]

I also encountered this annoying issue several times! In my experience, server preferably crashes, when a number of consolidated elements being referred to in several dimensions of the view to be extracted.

ibm states, that it still occurs even with HF6, see -->
http://www-01.ibm.com/support/docview.w ... wg1PM46816

"... TM1 server crashes when there are a number of write users working on a cube and a writer exports cube data with the Export as Text Data function."

ibm's recommendation is a 10.1 upgrade !!!

[holger_b]

... just an update re this:

It turned out that the problem occurs especially when the LockPagesInMemory parameter is present in the tm1s.cfg, no matter if it is set to true or false. Apparently this leads to problems especially if there is another database on the same server which uses that parameter as well.