Root Disk Online Replacement

[Arunkumar.B]

Hi All,

Superdome 16way in my setup having 9 vpars booting from 9 DS2120 Diskstations.Here we have encountered with frequent root disk failure.Checked Power Supplies & Environmental.Now HP has come up with solution to change all the remaining root disk proactively.



Servers are in production.How can i change the root disk online.

For Example

One of my vpar booted thru Primary Root disk.For changing the Alternate disk i can use lvreduce,vgreduce & pvcreate newdisk with normal mirroring procedure or else issue vgcfgrestore,vgchange,vgsync & mkboot.

But if i want to change my Primary rootdisk which is booted with primary root disk.I will have to shutdown & boot thru Alternate then give lvreduce..vgreduce... or Without shutdown & not booting thru Alternate can i go ahead with replacement

Cheers
Arunkumar.B

[Steven E. Protter]

Shalom,

Frequent root disk failure is not normally a disk problem. Its a problem with the other electronics and equipment that come with the disk.

If the root disk is mirrored, you should be able to pull the disk, which is hot swapable out and replace it without booting the system.

If its still responsive in any way, you should break vg00 mirrors and then pull the disk. Then follow the procedure you used to build the software mirrors.

If the mirroring was done right there is no reason you need to shut down the system for this.

SEP

[Patrick Wallek]

You can use the same procedure to replace your primary as you use for your secondary disk.

If you reduce the primary out of the VG, the secondary will take over.

[Patrice Le Guyader]

Hello,

When you say break the vg00 for replacing alternate , why not just simply stop lvm access on the "bad" disk (pvchange -a N pvname) ,do a vgcfgrestore on it,vgchange , do the mkboot,and finally the lvlnboot as describe in this whitepaper ?

http://docs.hp.com/en/7161/LVM_OLR_whitepaper.pdf

Am I wrong because I've done this on a vpar on similar configuration and didn't be able to boot on this alternate for testing (production server)?

Hope this helps
kenavo
pat

[Arunkumar.B]

Hello Steven,

This is Disks problem is confirmed by wtec-Australia says that G shock might be happend on the disks during transist.They have suggested to go ahead with proactive Replacement.

So can i change by this Manner as from your replies.
Primary Disk - c0t0d0
Alternate Disk - c0t2d0

Machine Booted with Primary.To change the secondry disk first.
#pvchange -a n /dev/dsk/c0t2d0
#vgcfgrestote -n /dev/vg00 /dev/rdsk/c0t2d0
#vgchange -a y vg00
#vgsync vg00
#mkboot -l /dev/rdsk/c0t2d0
#mkboot -a "hpux -lq" /dev/rdsk/c0t2d0



For Changing the primary disk which is booted thru primary.If i follow the same procedure will the server automatically will shift to Alternate during the time of replacement or i need to take a reboot using alternate disk


Regds
Arunkumar

[spex]

Hi Arunkumar,

Look at Fadia Almarei's post in the following thread for the procedure:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1000462

PCS