REPLACING DISKS

[MIGUEL CARABANO]

I need to replace two disks that estan bad in my Cluster atmosphere:
/dev/dsk/c2t1d0
/dev/dsk/c1t14d0

But it happens this:

file attachment:
Logical Volume To Physical Disk Mapping

That I must consider?

Greetings

Miguel C.

[Carlos Fernandez Riera]

I supose you are utilizando LVM mirror.

First ensure that all lv in those disks have his mirror in other disk.

lvdisplay -v /dev/vg.../lvxxxx | more

Then if you have hot plugable disks just extract faulty ones.

Insert new disks

run ioscan

restore lvm configuration:

vgcfgrestore /dev/vg... /dev/dsk/c.....

and syncronize


vgsync /dev/vg...



I think that is all.

[Rita C Workman]

I am not certain exactly what you are asking..but..from looking at attached file. It 'seems' you have 2 failed mirror disks. Your primary disks are still showing ok.
You don't mention the disk hardware..so I am just going to take a stab at this.
I might take down my packages involved..change vg so I have control of it..replace my bad disks...and since you don't mention your array hardware (EMC would resync) then you may need to resync either the lvol or vg. I might just go ahead and resync the vg to be on the safe side since that one disk covers a couple lv. But that's up to you. See manpage lvsync/vgsync.
===========================
cmhaltpkg -v
vgchange -c n /dev/vgname
vgchange -a y /dev/vgname
vgsync /dev/vgname
or
lvsync /dev/vgname/lvname
(Note a resync takes a good bit of time to complete..be patient..no bad messages means it is running..)
When done..
..Confirm everything looks ok with vgdisplay and lvdisply..

vgchange -a n /dev/vgname
vgchange -c y /dev/vgname
cmrunpkg -v
=============================
But wait for others, they may have noticed something I missed.

Rgrds,
Rit

[Patrick Wallek]

Attached is a document from the TKB (Doc ID - KBAN00000347) which outlines the procedures for replacing LVM disks in HP-UX 10.X and 11.X. It goes through all possibilities, mirrored boot disk, non-mirrored boot disk, mirrored data disk, non-mirrored data disk, etc.

[Sanjay_6]

Hi Miguel,

IF the disk is mirrored and how swappable, just remove the disk from the cabinet and then do a vgcfgrestore to restore the vg info to the disk. They do a vgchange -a e /dev/vg_name on the node where the disk is active. Then do a "vgsync /dev/vg_name" to sync the lv's for this disk. You should be fine after that. you don't have to shutdown the package / cluster or anything. If the disk is not mirrored or if the disk is not how-swappable, you have to shutdown the system. Shutdown the disk cabinet power, replace the disk, boot the system, start the cluster services on this cluster. do a vgcfgresote, activate the vg, and do vgsync.

Try this link,

http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/hpux/onlinedocs/B3936-90053/B3936-90053_top.html&con=/hpux/onlinedocs/B3936-90053/00/00/53-con.html&toc=/hpux/onlinedocs/B3936-90053/00/00/53-toc.html&searchterms=troubleshooting&queryid=20020222-072515

Hope this helps.

regds