Replace Primary ROOT Disk on ia64 hp server rx4640

[Mike_305]

Hello,

I need to replace the primary root disk on Itanium Server.

Q-1) Do I have to boot from Alternate or I can do this on ONLINE (On Fly)?

Q-2) And do I have to create the partition using IDISK command to create the partition before the mirroring or sync?

Do I still need to follow the normal process of VGSYNC like below.

vgcfgrestore -n /dev/vg00 /dev/rdsk/c1t5d0

mkboot /dev/rdsk/c1t5d0

mkboot -a "hpux -lq (;0) /stand/vmunix" /dev/rdsk/c1t5d0

vgchange -a y /dev/vg00

lvlnboot -Rv /dev/vg00

vgsync /dev/vg00

[Mike_305]

Sorry forgot to say

Appreciate your help.

Thanks,

Mike

[Robert-Jan Goossens_1]

Hi Mike,

Have a look at this complete disk faillure cookbook.

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

Regards,
Robert-Jan

[Mike_305]

Hello,

I already have this doc but what I am trying to find out is that this is primary disk and what happens on Itegrity/Itanium servers if I pull the primary disk?

And can I use vgsync to build the primary drive?

Regards,

Mike

[Sandy Chen]

Hi Mike,

Is your disk still available? If yes, safest way is you to boot from the alternate path, break the mirror and then re-mirror

If you are willing to take the risk, online is doable too. If you plug out the primary disk, it should go to secondary automatically, as long as your mirror configured properly.

And for Q-2, the question is yes, you have to use IDISK before vgcfgrestore.

regards,
Sandy

[Torsten.]

What server model is it?

Procedures could be different again on servers with SAS controllers!

BTW, disk devices for boot disks on Integrities always have a "s2" at the end for the data partition. Be careful - this is important (e.g. /dev/dsk/c1t5d0s2).

[Torsten.]

OK, it is a rx4640 - forget about the SAS.

But the procedures for 11.31 are different to the procedure for 11.23.

What OS version do you have?

[Torsten.]

OK, it is a rx4640 - forget about the SAS.

But the procedures for 11.31 are different to the procedure for 11.23.

What OS version do you have?

BTW, the steps you mentioned are for PA-RISC servers
(e.g. mkboot -a "hpux -lq (;0) /stand/vmunix" /dev/rdsk/c1t5d0 )

Do NOT use them on this Integrity!

[Mike_305]

Hello,

It is 11.23 running HP-UX.

Thanks,

M

[Tim Nelson]

Make sure that the vg thinks the disk is completely dead/unavailable and all LVs see that disk having stale mirror extents. Otherwise it will only sync the stale extends and leave holes.

If you can still talk to the disk I would completely reduce this disk from the lvs and vg, vgreduce it, swap the drive then follow the instructions just like you were adding a new vg00 mirror.

[Torsten.]

"I would completely reduce this disk from the lvs" - Tim this is not necessary since there is the "pvchange -a n ..." command (mentioned in the "when_good_disks_..." doc).

[Mike_305]

Hello,

May be last question.

One of my step is doing "Restore VG configuration to the replacement disk
# vgcfgrestore -n /dev/vg00 /dev/rdsk/c3t0d0s2

Do I have to do vgcfgbackup before vgcfgrestore OR this information will READ from ALTERNATE?

Thanks,

Mike

[Torsten.]

No need to do a vgcfgbackup now.

vgcfgrestore will read from

/etc/lvmconf/vg00.conf

see also

# man vgcfgrestore

[Mike_305]

Hello Everyone,

Thanks for help and appreciate very much.


Regards,

M