Background: SAP directory housekeeping is a common problem to basis engineers.
Below are some tips on how to handle problems with full or low space for usr/sap/SID.
Reason # 1.0: A huge file is unconsciously moved to usr/sap/SID directory.
Shit happens. Sometimes a person could accidentally move an ISO image or other unimportant file to a wrong directory.
This is very stupid but these things happen.
Solution # 1.0 Delete or move non-sap file.
Step 1.1 Use Windows Search tool.
Step 1.2 Search usr/sap/SID directory by file special option with 10Mb size.
Example of files that are non-sap related:
Installers (*.ins)
ISO images (*.iso)
unnecessary zip files (*.zip)
unnecessary rar files (*.rar)
Step 1.3 Verify if files are no longer needed.
Step 1.4 Delete or move the identified miscellaneous files
Step 1.5 It's always best practice to move instead of delete. Also Don't delete files WITHOUT second opinion of fellow system owners. Your choice.
Reason # 2: Standard SAP clean up jobs did not run.
Cleanup jobs must be regularly maintained for automatic SAP system cleanup.
Solution # 2: Verify if SAP standard reorg jobs below completed successfully.
a. SAP_REORG_JOBS
b. SAP_REORG_SPOOL
c. SAP_REORG_BATCHINPUT
See table below for more info on jobs:
In some urgent cases, you need to run programs associate with job manually. You can refer to my post below on how to manually run the programs associate with jobs above in link below:
http://basisbibo.blogspot.com/2009/09/fundamental-sap-system-clean-up.html
Sometimes SAP standard jobs like any other job will fail to complete. Troubleshoot if needed.
It's always best to check last entry of job logs. Some reasons of job failure could be:
a. Executing user is locked. Check user in SU01 and unlock if needed.
b. No background process available during scheduled runtime. This is temporary. Run the job again.
c. Disk I/O error. Check disk if full or physical errors are detected.
d. Variant does not exist. Create variant if needed.
Reason # 3.0: SAP_COLLECTOR_FOR_PERFMONITOR is not scheduled or having problems.
This job update the performance database MONI and reset stat file once the stat file reached it’s max size. This should run regularly.
Solution #3.0 Verify job status. Troubleshoot if needed.
Step 3.1 Go to SM37, enter SAP_COLLECTOR_FOR_PERFMONITOR and find job occurence.
Step 3.2 If the job does not exit, schedule it as follow in transaction SM36:
Job name: SAP_COLLECTOR_FOR_PERFMONITOR,
main program: RSCOLL00
Frequency: Hourly
No variant.
(Note: sometimes the job does not have the name of SAP_COLLECTOR_FOR_PERFMONITOR, the name depend on the job scheduler. You can check if it was scheduled in the system by the main program RSCOLL00 in SM37)
Step 3.3 Make sure that the job completes.
Reason # 4.0 Developer traces files(e.g. dev_w0, dev_w1, dev_w*) are huge.
Dialog, Background and Spool work processes logs are recorded and consumes space in disk as dev_w*.
Expect some work process to grow very large when major activity is ongoing like client copy or delete.
Solution # 4.0 Verify and reset dev traces if needed
Step 4.1 In OS level go to usr/sap/SID/DVEBMGS**/work
Step 4.2 Do a search for dev_w*. Verify if one dev_w* is getting huge size (Around 1Gb to 5Gb)
Step 4.3 In SAP level verify activity through transaction SM50. dev_w(No.field in SM50). Check the runtime if long running. 3600 = 1 hour.
Step 4.4 If confirmed to be long running, confirm with user if process can be reset. Don't reset process unless confirmation is given.
Step 4.5 With a green signal, in SM50 Click on Process > Trace > Rest > Work process files
Step 4.6 Verify in OS level if the dev_w* file is reset to 0~ kb.
Reason # 5.0 Very big stat file.
Stat files are regularly cleaned by job SAP_COLLECTOR_FOR_PERFMONITOR.
In some development systems, stat files are not really needed.
Solution # 5.0 Delete stat file
Step 5.1 In SAP level, go to transaction ST03
Step 5.2 Switch to Expert mode view
Step 5.3 Go to Collector and Performance DB
Step 5.4 Go to Statistics and Record file
Step 5.5 Go to Delete file
Step 5.6 Select the SAP system affected (server_SID_Instance no).
Reason # 6. Your SAP baby is getting big. :D
This is the last resort for system administrators for maintenance.
Most plugins are installed.
All clean up jobs okay.
No confirmation to reset stat or developer traces (dev_w*)
Solution # 6: Extend file system. This can be done through.
Step 6.1 Disk replacement if you have poor disk management infrastructure. This will required SAP downtime.
Step 6.2 Remote disk extension if you have Network File System.
0 Comments