HPUX upgrade backout plan - mirroring drives

[Michelle78]

I know this has been asked a thousand times, and unfortunately I do not feel comfortable applying the answers to my situation. Please take pity on me!

I have two mirrored OS drives in my hprp7420 B.11.11 system:

# vgcfgrestore -n /dev/vg00 -l
Volume Group Configuration information in "/etc/lvmconf/vg00.conf"
VG Name /dev/vg00
---- Physical volumes : 2 ----
/dev/rdsk/c0t6d0 (Bootable)
/dev/rdsk/c3t6d0 (Bootable)
#
I have a third drive at my disposal and will be upgrading from 11.11 to 11.23. I would like to create a backout plan by breaking the mirror and remirroring to the third drive, leaving me a bootable disk in case of a major problem.

Here is what I'm thinking:

Break the mirror, shutdown, yank a drive, add drive, rebuild (, shutdown, put original drive back, bring up and do upgrade. My hope is that if all goes to heck, I have backup drive to slap in, boot, and bring everything back up to pre-upgrade status.

Can someone help with commands and verify my shutdown locations, etc? I'm totally new on HPUX and the closest thing to an admin this place has, but i'm far from a unix person.

Thank you!

[Michal Kapalka (mikap)]

hi michelle78,

you procedure is ok, but before you start, make a Ignite/UX backup to the tape or ignite server. If somethings happend, you should be able with the ignite restore the system.


mikap

[Michelle78]

I knew I should have mentioned... ignite tapes go without saying.

I could also use some help with the actual command strings, but good to know I have the procedure down! Thx.

[Matti_Kurkela]

"Breaking" the mirror is slightly problematic here: lvsplit is not appropriate here, and lvreduce/vgreduce will leave the broken-off side in unbootable state (it is essentially marked as unallocated disk space).

Leave out the step of explicitly breaking the mirror at the beginning, and you should be good. (You'll also get a positive proof that you can boot your system with one-half of your mirror gone AWOL, which is good too.)

For instructions on rebuilding the mirror, please see:

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

The specific procedure for your PA-RISC server starts on page 48.

Note that the boot LIF at the very beginning of the disk is outside the LVM-mirrored area and must be replicated separately, using the "mkboot" command.

When creating a new mirror, _don't forget_ to use the -B option of pvcreate! Forgetting it makes the mirror non-bootable and the only way to fix it is to remove the mirror and do it over correctly.

Before beginning the procedure, make sure both your disks have the boot command set to "hpux -lq" (as opposed to just "hpux") so that your system will boot with no fuss even if any one half of your mirror is gone:

lifcp /dev/rdsk/c0t6d0:AUTO -
and
lifcp /dev/rdsk/c3t6d0:AUTO -

should both return "hpux -lq".
If one of them says just "hpux", fix it with:

mkboot -a "hpux -lq" /dev/rdsk/

After you upgrade to 11.23, you can use DRD which would make this entire procedure really easy to accomplish. Sadly, it is not available for 11.11.

https://h20392.www2.hp.com/portal/swdepot/displayProductInfo.do?productNumber=DynRootDisk

MK

[Michelle78]

Matti, Thank you so much for your help thus far. I've read the pdf suggested, and the comments you posted.

Here is the current system status:

# lifcp /dev/rdsk/c0t6d0:AUTO -
hpux
# lifcp /dev/rdsk/c3t6d0:AUTO -
hpux -lq
# ioscan /dev/dsk/c3t6d0
H/W Path Class Description
===========================================================
1/0/1/1/0/1/1.6.0 disk HP 73.4GST373454LC
# ioscan /dev/dsk/c0t6d0
H/W Path Class Description
=====================================================
1/0/0/3/0.6.0 disk HP 73.4GST373454LC
# pvdisplay -v /dev/dsk/c0t6d0 | grep 'current.*0000 $'
00000 current /dev/vg00/lvol1 00000
00019 current /dev/vg00/lvol2 00000
00530 current /dev/vg00/lvol3 00000
00549 current /dev/vg00/lvol4 00000
00551 current /dev/vg00/lvol5 00000
00926 current /dev/vg00/lvol6 00000
00939 current /dev/vg00/lvol7 00000
01314 current /dev/vg00/lvol8 00000
01814 current /dev/vg00/root 00000
01821 current /dev/vg00/misc 00000
01946 current /dev/vg00/users 00000
02071 current /dev/vg00/tmpdepot 00000
# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c0t6d0 (1/0/0/3/0.6.0) -- Boot Disk
/dev/dsk/c3t6d0 (1/0/1/1/0/1/1.6.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c0t6d0
/dev/dsk/c3t6d0
Root: lvol3 on: /dev/dsk/c0t6d0
/dev/dsk/c3t6d0
Swap: lvol2 on: /dev/dsk/c0t6d0
Dump: lvol2 on: /dev/dsk/c0t6d0, 0

Current path "/dev/dsk/c15t0d2" is an alternate link, skip.
Current path "/dev/dsk/c14t0d3" is an alternate link, skip.
Current path "/dev/dsk/c15t0d0" is an alternate link, skip.
Current path "/dev/dsk/c14t0d1" is an alternate link, skip.
Current path "/dev/dsk/c15t0d4" is an alternate link, skip.
# setboot
Primary bootpath : 1/0/0/3/0.6.0
Alternate bootpath : 1/0/1/1/0/1/1.6.0

Autoboot is ON (enabled)
Autosearch is ON (enabled)
#
# more /stand/bootconf
1 /dev/dsk/c0t6d0
#

Here is what I've written up as my procedures:

# mkboot -a "hpux -lq" /dev/rdsk/c0t6d0
shutdown the system and pull c3t6d0 drive.
Reboot system on one drive and verify it boots.
Shutdown the system and physically add spare drive.
Reboot system ----- the spare drive is pulled from another system. How do I keep from booting to it mistakenly?
Add device files: # insf -e -H 0/3/0.6.0
create physical volume on spare: # pvcreate -B /dev/rdsk/c3t6d0
add to root volume: # vgextend /dev/vg00 /dev/dsk/c3t6d0
add boot area: # mkboot -l /dev/rdsk/c3t6d0
check: # lifls -l /dev/rdsk/c3t6d0
add autoboot: # mkboot -a "hpux -lq" /dev/rdsk/c3t6d0
mirror volumes:
# lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol2 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol3 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol4 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol5 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol6 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol7 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/lvol8 /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/root /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/misc /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/users /dev/dsk/c3t6d0
# lvextend -m 1 /dev/vg00/tmpdepot /dev/dsk/c3t6d0

update root vg: # lvlnboot -R /dev/vg00
-------- is it a problem that right now lvol2 (which is not on bdf) is set for swap and dump and only on one drive???
set alt boot path: # setboot -a 0/3/0.6.0
add line to /stand/bootconf ----- is it a problem that c3t6d0 is not in this bootconf file at this time?

shutdown system and pull c3t6d0 drive again, replace with original c3t6d0 drive and perform upgrade. Hope and pray I don't need the spare I've created.


What did I miss?

Thanks!