MCSG: Resizing (Reducing) A Shared File System

[LG Porter]

I have a two node cluster running MCSG 10.10 and HPUX 10.20. This cluster has a old version of Adv.Online JFS. How is this best performed? That is there have been issues with older versions of Online JFS. The file system is presently 25GB (25000Megs)and needs to be reduce to 10GB. Traditionally two commands would be used to dynamically perform this:
lvreduce -L 10000 /dev/vg01/lvol1
fsadm -F vxfs -b 10000M /v1 (v1 is the file system mount point. Also some syntax uses the block size or I've been told by HP Reponse Ctr Engineers that the "M" option could also be used referencing megs with the fsadm command.) If the fsadm command fails to dynamically resize the file system, would the file system have to be taken off line and a "newfs" used? Tahnk You.

[James R. Ferguson · Accepted solution]

Hi:

The general result of trying to reduce the size of a VxFS filesystem prior to Online JFS version 3.3 (which is not available for 10.20) is failure.

You can try defragmenting the filesystem before attempting to reduce it's size but this will not guarantee success in a subsequent size reduction.

You need to seriously consider the long route of backing up your data and doing a 'newfs' of the filesystem as you stated.

Regards!

...JRF...

[Peter Kloetgen]

Hi,

as James allready told, the old online JFS versions fail if you try to reduce a file system. So do a backup before you try it!!!

Allways stay on the bright side of life!

Peter