lvm

[irshad_1]

i am using hp 9000 D class server with (oracle Data base ) in HP-UX 10.2.this server has two volume group.this are vg00 and vg01,one of lv in second volume group
disk has some bad blocks ,we are going to replace disk (same Size).now able to read the data from the disk,
can any one help me with procedure to do the activity?

[Manix]

You have 10.20 pretty old OS & you are going to replace the disk ,i am not sure if you can do this on line on this version.

After replacing the disk you need to run
vgcfgrestore -n vg01 /dev/rdsk/cxtxdx & then backup your data using mirror or any other source.

Thanks

Manix

[Torsten.]

The "official" procedure:

- unmirror the disk
- swap the disk
- mirror the disk

steps are here:

When_Good_Disks_Go_Bad_WP
http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01911837/c01911837.pdf

[Emil Velez · Accepted solution]

forgot remove the disk from the volume group in the above procedure.

Otherwise good. By removing the disk from the volume group you can add a larger disk up to max pv per pv to your volume group like removing a 2GB disk and then adding a 4 GB disk to the volume group.

You could also try

Add new disk to volume group
pvmove from old disk to new disk
vgreduce old disk
do screwdriver work

Assumptions above assume old disk is still somewhat operational.

[P Arumugavel]

If you could able to read the data from the disk, then you can go for pvmove.

Add the disk in the volume group
Do the pvmove from disk to disk
Then remove the disk from the volume group.

#pvcreate newdisk
#vgextend vgname newdisk
#pvmove olddisk newdisk
#vgreduce vgname olddisk

Note that pvmove command will fail if LV is striped.

Rgds...