Taken from http://madhutekur.wordpress.com/2008/08/17/housekeeping-on-sapmntglobal/
While I was performing my daily monitoring tasks, I found that many Jobs had got cancelled, and when I tried to read the job log to find out what was the error I couldn’t read any of the logs. I received an error “Error reading Job log”.
When I checked the System log (SM21), the logs read
“Failed to create log for job <abcde> 04534901%_EVENT
Error synchronizing file /usr/sap/<SID>/SYS/global/100JOBLG/0001X045349”
When I checked the /sapmnt/<SID>/global directory, I found that
100% full. This was the reason why the jobs were failing; the job was not able to create its job log.
Due to this file system full, many BW process chains also got cancelled and lots of tickets were getting raised. Just then realized how important it is monitor file system when we perform our daily checks.
Now let me tell how housekeeping can be done on this file system /sapmnt/<SID>/global. The global directory contains the job logs. Old job logs can be deleted by running report RSBTCDEL2 in se38. Choose the number of days old logs that can be deleted. We must make sure that the standard SAP job SAP_REORG_JOBS is run, so that old job logs get deleted automatically.
The global directory also contains some files of format WF_LOGS*. These are work flow trace files. The note 113656 gives us information about these logs. You can use transaction SWU10 or SWT0 transactions to delete the old work flow trace files.
In general, you can schedule a job that will delete the logs or just use transaction SWU10 -> ‘Delete obsolete workflow trace files’. To schedule, just go to the menu: Program => Execute in background and schedule from there. The report that it is running is RSWTTR02, so if you have any problems with SWU10, you can always schedule the job manually via SM36.
No comments:
Post a Comment