Generation of large number of archive logs

[Former Member]

Hi,

On my CRM production system, BBPCRM4.0 on HPUX 11.23 and Oracle 9i. We have applied the support packages for ABAP and BASIS upto level 63 and for the BBPCRM4.0 upto level 13.

After doing this activity we have found that the number of redo-logs getting gererated, We have scheduled extra online archive log backups....

We have referred to note number 581359, but after implementing the same, the log generation reduced,but the transactions being executred were getting timed out and the system was very slow, so we reverted back to the earlier parameter

The log generaton is at least 10-12 logs of 50 MB per minute. If any one comes across any solution to this please revert back.

Thanks in advance

Regards

Alfred

[stefan_koehler]

Hello Alfred,

without knowing something about the load / users on your system.. but i don't think that you run in a oracle bug problem after your SAP SP upgrade.

>After doing this activity we have found that the number of redo-logs getting generated

Is your database is maybe still in backup mode?

Have you already run the SGEN after your SP upgrade?

Maybe the generation of the "new" programs is responsible for the redo log generation.

If not take a look at the oracle sessions (ST04 -> Detailed analysis -> Oracle Sessions) and have a look at some long running DML / DDL statements.

If all this activities ares not pointing to something, you can also use the log miner to take a look into the redolog files and check what is filling the redo logs.

LogMiner: http://download.oracle.com/docs/cd/B19306_01/server.102/b14215/logminer.htm

Regards

Stefan

[markus_doehr2]

If you install support packages those are changes in the database and depending on the number of those this will heavily increase your archive logs. This is normal and expected behaviour.

Reg. the note: are you on that low patchlevel to apply the note?

Markus

[Former Member]

Hi Markus,

We have refered to the statement within the SAP Note - ALTER SYSTEM SET fast_start_parallel_rollback = false;

After making this change the system started creating the logs in a normal manner. We executed the statements in order to arrest the creation of the log files.

One more thing that has come to my mind is the PSAPUNDO has no space Given below are the details

List of tablespaces for extension

Pos. Tablespace Files/AuExt. Total[KB] Used[%] Free[KB] ExtSize[KB]

1 - PSAPDAT 200/200 787865600 94.95 39801280 113561600

2 - PSAPDAT620 5/5 11407360 80.61 2211456 28262400

3 - PSAPDATUSR 1/1 1048576 6.52 980224 9191424

4 - PSAPECPDB 1/1 1024000 28.56 731584 9216000

5 - PSAPTEMP 6/2 28868608 2.12 28255232 0

6 - PSAPUNDO 18/7 57589760 100.00 0 0

7 - SYSTEM 1/1 1048576 31.93 713792 9191424

I agree that after a support pack there will be a huge number of archive logs, I executed the SGEN and it got completed...We completed the activity at around 12:30 am then the SGEN was scheduled immediately and that got over by by around 1:30am. After that all the online backup jobs that were fired completed successfully. The last backup schedule was at 8:30AM which completed successfully.

After that the problem has started

These are just some details that I've put regarding the issue.

Alfred