1832151 Members
3080 Online
110038 Solutions
New Discussion

Disk problem

 
SOLVED
Go to solution
Otto Folprecht
Advisor

Disk problem

Need help. One of our disk's died on us so we replaced the disk under warranty. Since the system is hot swap I replaced the disk and tried to mount the disk with mount /u3. The response I got was mount: /u3: I/O error. In sam when trying to access the disk I get the same answer. So what do I have to do to get this disk back online. It's not mirrored or anything to that extent. The disk is solely vg04 on lvolu3 (dev/vg04/lvolu3). THX Otto
7 REPLIES 7
Steven E. Protter
Exalted Contributor

Re: Disk problem

slow down.

insf -e

pvcreate the disk.

You may need to boot before any of this works.

After pvcreate you should be able to get it back into a volume groups and synch mirrors if it originally was mirrored.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Jeff Schussele
Honored Contributor

Re: Disk problem

Hi Otto,

You need to restore the VG info to it first:

vgcfgrestore -n /dev/vg04 /dev/rdsk/cXtYdZ

This assumes that you ran vgcfgbackup on vg04 - but that happens by default on a lot of vg commands.

Then

vgchange -a y /dev/vg04

to activate it - then mount the *all* the LVs that comprised vg04. At that point IF they mounted OK - you can restore to them from tape.

As Tully stated - slow down - it takes a few steps to get to your destination. Hot-swap doesn't mean spare disk allocation.

Rgds,
Jeff

PERSEVERANCE -- Remember, whatever does not kill you only makes you stronger!
Patrick Wallek
Honored Contributor

Re: Disk problem

Even though the disk has been replaced, it has NO volume group structures on it. So you need to get those back first.

# vgcfgrestore -n /dev/vg04 /dev/dsk/c?t?d0

I think the syntax is correct. Check 'man vgcfgrestore' to be sure.

Once that is done, you can do:

# vgchange -a y /dev/vg04
# /u3

Now you will need to restore your data.
Patrick Wallek
Honored Contributor

Re: Disk problem

You might also want to take a look at the "HP Software Recovery Handbook" located here:

http://www2.itrc.hp.com/service/iv/node.do?node=prodITRC%2FWW_Start%2FN1%7C16

Specifically take a look at Chapter 16 which deals with LVM. There are very good details and instructions on how to replace disks.

Shaikh Imran
Honored Contributor
Solution

Re: Disk problem

Hi,
Please do the following:
Install the device file for the disk by
1) # insf -e
Just get the device file of the disk by
2) # ioscan -fnC disk
say the device file is cXtXdX.
3) #pvcreate -f /dev/rdsk/cXtXdX
4) #vgcfgrestore -n /dev/vg04 /dev/dsk/cXtXdX
5) #vgchange -a y /dev/vg04
6) #mount /u3

By any chance if you are not able to restore the VG configuration by vgcfgrestore then you can create a different vg e.g. vg07 and also change the mount point in the /etc/fstab file.

Now restore the data.

Regards,







I'll sleep when i am dead.
Kent Ostby
Honored Contributor

Re: Disk problem

Otto -- If this disk is a mirrored disk, the vgchange -a y /dev/vg04 may hang for a while.

This is normal as the mirroring is resynced.

If its not a mirror and is a filesystem then you will need to redo the FS structures between steps 5 and 6 of Shaikh's steps above using fsadm or newfs depending on your filesystem type.

Best regards,

Kent M. Ostby
"Well, actually, she is a rocket scientist" -- Steve Martin in "Roxanne"
Trond Haugen
Honored Contributor

Re: Disk problem

HotSwap means you can remove and incert a disk without having to power off the disk cabinet. furhter more HP-UX will hadle this (meaning it will see a device "go" and "come").
BUT data on the disk will only be restored if you have mirroring. So as you don't have that all you will have after swapping is a disk with littery nothing on. (No LVM header, no vg, no lvol, no data.)
Doing a vgcfgrestore will put the LVM information back on the disk. and you will be able to mount it. Then you can restore the data.

Regards,
Trond
Regards,
Trond Haugen
LinkedIn