We are in a Project to move from Oracle (10.2) to DB2 (DB6 Ver 09.07?) on our ECC/BW/APO systems..
We have Live still operating on Oracle, and QA system (copy of Live then converted) running on DB2.
We have an InfoCube in BW (holds versions of our Entire Demand Plan, one version per day!) which has a lot of records (too many really but thats another story!!!!), in our Live (Oracle) Overnight Process Chain this takes just short of 30 minutes to re-add Indexes after data is loaded, but in DB2 environment this is taking over 10 hours (even on a system with no other users but me it took 8.5 hours!). This is the same every day, so not just a first run type of issue.
Note: I am no expert at Db level !!!!!
The cube is marked as "Index Clustering", as opposed to Multi-Dimensional Clustering, and I see I can't change this as I have data loaded).
There are no specific messages showing in SM21 or ST22 that I can see, Nothing shows in Job Log.
There is 2 years worth of data in the cube (yes I know this should not really be the case but I am new to organization and it seems to historically be this way, I need to rework the entire data model, but that will have to be after this project!).
I have a High call with SAP open but it has been 4 days with no real response and they won'tr work on it over the weekend apparently, but I need to get this moving otherwise Project Go-Live is threatened!
Are there any gurus out there who can point me in the direction of what might be the issue, or what to check, and of course how you might go about getting this back to normal runtime? Our overnight processes would not complete anythwre near to the required time if they went for an extra 10 hours!!!!!
Regards
Simon