I have a job where I load five hashed files from five Oracle tables. There are ten to twenty of these jobs. Lately I have stated to experience problems with these jobs failing with following error messages. I tried to reindex DS_STAGETYPES as I though it might have something to do with that but that didn't help. Problematic with this is that it happens almost every day but with different jobs - although they are similar and the error messages are always the same. When I compile the job and run it again everything usually works fine.
Any ideas what might be causing this kind of behaviour?
Program "DSR_GETPROP": Line 63, Error initializing AK file "H?".
Job Aborted after Fatal Error logged.
Attempting to Cleanup after ABORT raised in stage D_Product_prepare..MaxKey.IDENT1
DataStage Phantom Aborting with @ABORT.CODE = 1
Program "DSR_GETPROP": Line 63, Error initializing AK file "D:/CAB_DW_PROD_10/I_DS_STAGETYPES/INDEX.001".
Job Aborted after Fatal Error logged.
Attempting to Cleanup after ABORT raised in stage D_Product_prepare..ProductGroupHash.IDENT3
DataStage Phantom Aborting with @ABORT.CODE = 1
'Pin 1 not initialized' Here the Pin 1 is pointing to one end of the link 1 to the stage IDENT3.
Try to clear the hashed file and rerun. Also make sure you are running the job with the same user id.
Impossible doesn't mean 'it is not possible' actually means... 'NOBODY HAS DONE IT SO FAR'
The good news is that the error doesn't have anything to do with your job, it is a DataStage engine issue. The initial messages point to a corrupt index on the DS_STAGETYPES hashed file. I just checked to make sure and the 3 indices on this table should be fixed with a DS.REINDEX.
If you try the command "DS.REINDEX DS_STAGETYPES" from the TCL or Admin command window you will get a screenful or two of information, are there any error messages or warning when you do this? It should show you the index list at the end and all 3 indices should have a build status of "Not Reqd"
I was also thinking the problem must be with DS_STAGETYPES as this is always referred to in error messages. I ran DS.REINDEX and it went smoothly - no warnings and all three indices with build status Not Reqd. And worst of all - the problem didn't go anywhere (well it emerged in a different job this time )
It certainly seems like Engine issue. Could this kind of behaviour be caused by some disk problems? Or other hardware related issue? This puzzles me...
Now we get similar error stating problems with DS_JOBOBJECTS. I'm starting to think that our project is about to crash... Maybe I'll try to create a new project and move jobs there.
Before going anywhere...
First BACKUP all your exising project.
Do a full export of your project and also do a unix level back up.
From TCL check the pointer strength with COUNT DS_JOBOBJECTS and make sure you are in safer side.
Impossible doesn't mean 'it is not possible' actually means... 'NOBODY HAS DONE IT SO FAR'
Read my previous post. You have at least one corrupted index on DS_STAGETYPES. The file D:/CAB_DW_PROD_10/I_DS_STAGETYPES/INDEX.001 (mentioned in the error message) is the first index defined for DS_STAGETYPES.
Rebuild at least the DS_STAGETYPES indexing.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
just to post the cause of the problem: Anti virus software
Yes, I know it should be obvious to check this out first but we had done the same exercise couple of months ago with same kind of mysterious problems. The end result back then was not to include project folder in the virus scan. And jihaa! after that everything worked... Until this problem emerged. Now we just double checked that the project folders were still excluded from the scan and to everybody's surprise someone had turned the scan back on in those folders.
Lesson: Don't trust in that anything stays the same over time - always check that what you think is happening actually is happening!