replace non-root mirrored jamaica disk ? Do I need to reduce lvreduce all vols ?

[Sammy_2]

Do i really need to lvreduce all vols on the bad disk and vgreduce the disk before I do
the steps below so there is no i/o OR

Are these steps correct and
can they be done all online on 10.20 system ?

=====================
# pull out c0t8d0 from array
# vgcfgrestore -n /dev/vg02 /dev/dsk/c0t8d0
# vgchange -a y /dev/vg02
# vgsync /dev/vg02
Is that it ?
===================





# uname -a
HP-UX inncux14 B.10.20 A 9000/800 1866711461 two-user license
inncux14:root# model
9000/800/K380

# pvdisplay -v /dev/dsk/c0t8d0|more
--- Physical volumes ---
PV Name /dev/dsk/c0t8d0
VG Name /dev/vg02
PV Status unavailable
Allocatable yes
VGDA 2
Cur LV 1
PE Size (Mbytes) 4
Total PE 4340
Free PE 0
Allocated PE 4340
Stale PE 1267
IO Timeout (Seconds) default

--- Distribution of physical volume ---
LV Name LE of LV PE for LV
/dev/vg02/dmadminlv 4340 4340

--- Physical extents ---
PE Status LV LE
0000 stale /dev/vg02/dmadminlv 1250
0001 stale /dev/vg02/dmadminlv 1250
0002 stale /dev/vg02/dmadminlv 1250
0003 stale /dev/vg02/dmadminlv 1250
0004 current /dev/vg02/dmadminlv 1250
0005 current /dev/vg02/dmadminlv 1250
0006 stale /dev/vg02/dma

[Patrick Wallek · Accepted solution]

No, no need to lvreduce / vgreduce. I never do that when replacing a disk.

Your steps look great!

[Devender Khatana]

Hi,

There is no need to reduce the LV/VG from the defective disk before the replacement but you can stop LVM access to that disk.

But in some scenarios you may be required to do that. The official guide explains this here.

http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad.pdf

HTH,
Devender

[Sammy_2]

That is exactly what I did and it worked fine. I was little skeptical of too much i/o going to bad disk (which had not failed completely but had some stale extents.
Dev and Pat, Thanks to both of you (Dev and pat)

[Sammy_2]

Dev and Pat provided perfect solution and Dev had a handy guide as well.