cancel
Showing results for 
Search instead for 
Did you mean: 

LOVs in Universe migrated to BI 4.1 using UMT

blair_towe2
Participant
0 Kudos
146

Good morning - We are (finally!) starting the migration away from Deski for our operational reports. We have a BOBJ XI 3.1 SP 7 FP 7.2 system running on Windows and we have a BI Platform 4.1 SP 6 Patch 1 system that runs on AIX. We are using the Upgrade Management Tool to bring the 3.1x universes over to the 4.1 system. I installed a 4.1 SP 6 Patch 1 system on Windows to act as the UMT host. The universes are coming over with no errors and they can be opened in Universe Builder with no issues.

The issue we are seeing is that after the universe is imported to the BI 4.1 system, it looks like it is then trying to fill any LOVs that exist on the universe and we are experiencing performance problems on the source database because of this. The source database is an Oracle 11.2.0.4 DB running on AIX. We are pretty sure that it is LOVs being filled that is causing the extra load on the DB because when we look at the processes, we can see that they are select statements on fields where there are LOVs setup up on the universe.

Has anyone else experienced these type of performance problems, and is there any way short of deleting the universe in BI 4.1, deleting the LOVs in the BOBJ 3.1 system, and then doing the migration to 4.1 again? The LOVs are still needed, so this would mean a lot of rework.

Is there a way to specify a time when to kick off the processes to fill the LOVs or to limit how many processes/jobs that are used to do this work?

Your assistance is appreciated!

Regards,

Blair Towe

View Entire Topic
denis_konovalov
Active Contributor
0 Kudos

I'm pretty sure what you're seeing is not related to LOV's, but related to Platform Search.
You need to review following KBA https://service.sap.com/sap/support/notes/2241459

blair_towe2
Participant
0 Kudos

Thank you Denis! It appears this was the issue. I have changed the parameters on the Platform Search Server and we are no longer experiencing the issue. We previously had the settings set to do a continuous crawl for all platform content. We changed this to collect Platform and Document Metadata and now when we migrate a universe we no longer see the continuous hits on the target database like we did before.

Regards,

Blair Towe