File system full sap in unix AIX

1. Check work directory for core dumps and log files with extension old.

Path : /usr/sap/<SID>/<instance>/work

rm *.old

rm core

2. If you find any huge file other than .old and java core file check the date last updated and nullify that file.

cp /dev/null <file name>

3. Check /usr/sap/<SID>/<instance>/j2ee/cluster/server0  directory for java core files.

If there are 2 nodes then you need you check for server1 also.

4.Check archive directory for old log files. (*.txt and *.zip files)

find /usr/sap/<SID>/<instance>/j2ee/cluster/server0/log/archive -name "*.txt" -mtime +2 | xargs ls

find /usr/sap/<SID>/<instance>/j2ee/cluster/server0/log/archive -name "*.zip" -mtime +2 | xargs ls

remove if found any files from the above command.

find /usr/sap/<SID>/<instance>/j2ee/cluster/server0/log/archive -name "*.txt" -mtime +2 | xargs rm

find /usr/sap/<SID>/<instance>/j2ee/cluster/server0/log/archive -name "*.zip" -mtime +2 | xargs rm

 find . -mtime -7 -type f -size +2000K

If you are unable to free up disk space in dev or qa, you can nullify the traces and close the ticket. Don't ever try this on production.

Example:

cd /usr/sap/SID/DVEBMGS02/j2ee/cluster/server0/log

find . -name "*.trc" -exec cp /dev/null {} \;

find . -name "*.log" -exec cp /dev/null {} \;

find /usr/sap/SID/DVEBMGS02/j2ee/cluster/server0/log/archive -name "*.zip" -mtime +2 | xargs ls

find /usr/sap/SID/DVEBMGS02/j2ee/cluster/server0/log/archive -name "*.zip" -mtime +2 | xargs rm

find /usr/sap/<SID>/<instance>/j2ee/cluster/server0/log/archive -name "*.zip" -mtime +2 | xargs rm