[HOWTO] file systems and tune2fs

DanGiles

ADSM.ORG Senior Member
Joined
Oct 25, 2002
Messages
626
Reaction score
17
Points
0
Location
Toronto, Ont. Canada
PREDATAR Control23

Interesting I haven't seen this issue posted already.
We are running TSM 8.1 on RHEL 6.7, but this should apply to any TSM server version, and most UNIX.
What do people do around setting up their LUNs and automatic disk checking? When you have a couple 100 TB setting at the end of your cables, reboot could take hours. So I am thinking I should disable e2fsck on all of the storage pool volumes, but keep it on the db and actlog volumes.
What do other people do?
 
PREDATAR Control23

That is what I do - fsck (regardless of version) takes a long time to run for big volumes.

This will be different if you have smaller volumes; fsck runs faster but the whole process still takes time. It would be wise to run fsck once in a while.
 
PREDATAR Control23

I wouldn't run fsck when TSM is running - the reason is that it may want to fixed a corrupted block and TSM would see as a grab-and-hold of that section of the storage pool and mark that as unavailable.
 
PREDATAR Control23

Hmmm. Do you schedule down time, or do you have some sort of boot script that would pick out one file system and run fsck on that one only, cycling through each one? I've got 35 4TB volumes!
 
PREDATAR Control23

Hmmm. Do you schedule down time, or do you have some sort of boot script that would pick out one file system and run fsck on that one only, cycling through each one? I've got 35 4TB volumes!

I do schedule down time - a necessary evil.

However, this is not that bad with AIX as with Linux. In AIX, the system can run for much longer without doing a warn restart for file checks to run.

My AIX systems used to reboot every 10 months were as my Linux ones are every 180 days or 6 months. I find that AIX is more resilient against disk corruption.
 
Top