on 01-26-2016 2:35 PM
Our application log is getting so large that it is causing performance problems.
I noticed that we have never archived or reorganized our application log.
I know that there are 2 options for managing the application log.
Which option do you use at your company?
I tried testing the archive functionality in our QA test system and it was very slow.
It took over 7 hours to delete the following number of entries.
Type No. Description
BALDAT 293,908 Application Log: Log data
BALHDR 61,254 Application log: log header
BAL_INDX 13 Application Log: INDX tables
Has anyone else encountered such poor performance?
QA systems often do not have the power of a production system, so it could be much faster there.
7 hours for an archiving job would not give me any headache. Especially with an initial archiving I used to run times from several days to several weeks.
For the application log we use a mixed approach, our EHS department wants a certain object be archived, all others will be deleted.
We had 800 million records in BALDAT, among them countless orphaned (there are OSS notes available). It took weeks to get this down to a reasonable level, since then the deletion and archiving jobs run weekly.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Our QA system is a copy of PRD and run times between the 2 have not been too far apart.
However, 7 hours to delete the number of records shown in my original message is horrible.
In the most recent years, we has 2-3M records per year in BALHDR. That would take forever to run.
We have been live since 1999 and never did any archiving until I started the project last year.
I would probably need to implement a hybrid approach too.
I would purge all logs more than x years old and then decide by object type what should be archived versus deleted.
All SAP archiving programs supposed to be developed with one limitation: that they occupy small space in RAM (e.g. 100MB) while in execution so if they do not finish over night or over weekend, they would not bring a SAP server to its knees during business hours. So in this case it is good they are slow, because you can let them run during business hours without problems.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
84 | |
25 | |
12 | |
9 | |
6 | |
6 | |
5 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.