New Disk Error

[Nobody's Hero]

I had a primary internal disk on My Nclass die. Went on chugging off of the mirror with no interuption. The bad disk actually went into a "NO_HW" stat in ioscan.

SOOoo, I replaced the disk while the server was running. No problem. When I tried to extend vg00 to the new disk I got an i/o error.

Soooo, I looked in syslog and found this:
HP has never seen this before. I have a case open. Has anyone seen this before?

The part number is OK..
0/0/2/0.6 tgt
0/0/2/0.6.0 sdisk
LVM: VG 64 0x000000: PVLink 31 0x016000 Wrong Disk!

[James R. Ferguson]

Hi Robert:

You have to do a wee bit more than just hot-swap the disk. See here:

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

Regards!

...JRF...

[Patrick Wallek]

What steps did you do after you replaced the disk?

Did you force the removal of the disk from the VG before you replaced it?

[Geoff Wild]

Shouldn't have to vgextend...try:

(assuming the disk is c1t5d0):

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

Rgds...Geoff

[Nobody's Hero]

When good disks go bad"

The way I read it.
If the disk is in a "NP_HW" state, then you can swap it. So I did.

[Nobody's Hero]

When good disks go bad"

The way I read it.
If the disk is in a "NO_HW" state, then you can swap it. So I did.

[Nobody's Hero]

Geoff,

Can you re-reply. You fixed it. I need to grant additional points....

Thanks.

[Nobody's Hero]

If its a new disk. Why wouldn't you have to extend the vg onto the new disk?

I don't understand.

[Pete Randall]

That's what the vgcfgrestore is for. It restores the LVM configuration from backup, negating any need to manually do all that stuff!


Pete

[Torsten.]

Robert, all the disk related information is stored on your filesystem.

If you do a vgcfgrestore, you re-create the lvm structure on the "new" (replaced) disk.

If you do a pvcreate + vgextend, you have the information about the old disk AND the new disk in the kernel/LVM. You will get LVM errors, at least if you do something like vgcfgbackup.

So if you swp a disk, just do a vgcfgrestore to this disk and inform the kernel (vgchange -a y vgxx), restore the boot information if needed and sync - thats all.

[Torsten.]

The error could look like this:

vgcfgbackup: /etc/lvmtab is out of date with the running kernel:Kernel indicates
3 disks for "/dev/vg03"; /etc/lvmtab has 2 disks.
Cannot proceed with backup.


Resolve it by typing a

# vgreduce -f vg03
PV with key 2 sucessfully deleted from vg vg03
Repair done, please do the following steps.....:
1. save /etc/lvmtab to another file
2. remove /etc/lvmtab
3. use vgscan(1m) -v to re-create /etc/lvmtab
4. NOW use vgcfgbackup(1m) to save the LVM setup


HTH!

[Geoff Wild · Accepted solution]

Robert,

Here's my 2 cents on bad disks:

1) If the disk is completely "dead", such as if you run ioscan and status is "no_hw" then you can hot swap the disk online.

2) However in circumstances where the disk has not fully failed please do one of the following to avoid data corruption :

a) reduce mirror before replacing the disk
b) deactivate VG before replacing the disk
c) shutdown system to replace the disk


Glad it worked!`

Rgds...Geoff

[Sameer_Nirmal]

Hi,

The "New Disk Error" reported in our case is a part of LVM produt change done for error messages given by LVM. The change was done for HP-UX 11.00 onwards through a patch.

The earlier version of this error was
LVM: Failed to restore PV nn to VG mm! Identifier mismatch."


Since you changed the disk and doesn't have any LVM structure present on it, the error is being reported. The LVM structure would get created using "pvcreate" command making the disk as "LVM disk" so LVM can use it.

Using "vgcfgrestore" command restores the VG information alongwith the LVM structure on the disk making the disk as part of the VG thereafter. You just need to do the step as mentioned by Geoff.

[Shameer.V.A]

Hi Robert,
Please find the attached document,

Regards,

Shameer

[Devender Khatana]

Hi,

As the disk gave I/O error while remirroring, I would suspect the new disk itself and would go for a read/write test on the disk using dd before proceeding further on mirroring part. By doing so atleast you will be sure that atleast the disk is allright and will see other details.

Otherwise also in scenarios where you see multiple reasons for a problem, just eliminate them one by one rather then fighting with all of them at the same time.

dd read test

#dd if=/dev/rdsk/cxtydz of=/dev/null bs=1024k

dd write test

#dd if=/dev/null of=/dev/rdsk/cxtydz bs=1024k

Caution - dd write will delete everything on the disk so if is already a part of some configuration then remove it.

HTH,
Devender

HTH,
Devender

HTH