cancel
Showing results for 
Search instead for 
Did you mean: 

DB performance in RSA1-process of repaire DB indexes too long n failed

Former Member
0 Kudos

Hi All,

While using the standard functionality in RSA1 to update / repair DB indexes for a Profitability Analysis infocube, 3 steps needs to be follow:

1. Check DB indexes

2. Delete DB indexes

3. Repair DB indexes

The problem comes when the system tries to delete / repair the indexes whereby it will run too long and finally, we need to "kill" the process by cancelling it. This happens in QA server. While testing it in Development server, it runs well.

I have scan the SQL statement behind the 3 process above using tcode SM50 and submit these statements via ABAP program withing EXEC SQL ..... ENDEXEC. This program, when executed in QA server it will still gives a long execution time. When one of the index won't be able to delete, the next step (repair) will fail as well.

What is the reason of the above issue? Can anyone help?

Thank you.

Accepted Solutions (0)

Answers (1)

Answers (1)

markus_doehr2
Active Contributor
0 Kudos

- Do you have the same amount of entries in the table on DEV and QA?

- Do you have identical hardware (OS, I/O subsystem etc.)?

- Do you have the same patches installed?

- Same parameters on both databases?

Markus

Former Member
0 Kudos

Hi Marcus,

Thanks for the reply. When I checked for the question you have posted above, I have found that the patches is not the same between both DEV and QA. Hopefully after updating the patches, we can delete and repair the DB index using standard code RSA1. Will keep you updated on the results.

Thanks.

markus_doehr2
Active Contributor
0 Kudos

Ok!

The first thing I would have checked is the number of table entries. An index repair of a table with 10,000,000 entries takes longer than the same on just 1,000,000.

Markus