Virtual TapeServer 6.04.01 Operations and Administration Guide

196 |
p. Upgrade the GFS file system:
WARNING Downgrading the GFS environment after this step requires rebuilding
the GFS cluster as well as all user data.
• Determine the GFS volume group names in vgscan by entering this command:
vgscan
Here is an example of the output:
Reading all physical volumes. This may take a while...
Found volume group "pool_gfs01" using metadata type pool
Found volume group "pool_ccs" using metadata type pool
• For each GFS volume group, start the upgrade process, replacing gfs_pool with
the GFS volume groups from the vgscan output. Do not run these commands
against the ccs volume group (such as pool_ccs).
vgconvert -M2 gfs_pool
vgchange -ay
• Create a mount point for GFS mounts:
mkdir -p /mnt/gfs
• For each GFS volume, mount the volume and change the lock manager:
mount -t gfs -o upgrade /dev/gfs_pool_dir/gfs_pool /mnt/gfs
umount /mnt/gfs
gfs_tool sb /dev/pool_gfs/gfs_pool proto lock_dlm
When prompted whether you are sure, enter y.
WARNING
Failure to enter the gfs_tool command will result in the following
error when mounting the GFS file systems:
/sbin/mount.gfs: error 2 mounting devicename on
mountpoint
Here is an example for a system that has two GFS pools (pool01_gfs and
pool02_gfs):
mount -t gfs -o upgrade /dev/pool_gfs/pool01_gfs /mnt/gfs
umount /mnt/gfs
gfs_tool sb /dev/pool_gfs/pool01_gfs proto lock_dlm
Are you sure? [y/n] y
current lock protocol name = "lock_gulm"
new lock protocol name = "lock_dlm"
Done
mount -t gfs -o upgrade /dev/pool_gfs/pool02_gfs /mnt/gfs
umount /mnt/gfs
gfs_tool sb /dev/pool_gfs/pool02_gfs proto lock_dlm
Are you sure? [y/n] y
current lock protocol name = "lock_gulm"
new lock protocol name = "lock_dlm"
Done