disk name changed

[Kaleung Tsou]

Hello,

I am running HP-UX 11.23 with the system connected to EVA 5000. This morning a HP service person did some maintenance on the SAN switch and when I boot my system, the disk name changed and the system is not able to mount the disks.

The original disk names are below by running strings /etc/lvmtab

/dev/vg00
/dev/dsk/c2t1d0s2
/dev/dsk/c2t0d0
/dev/vg02
dKF1
/dev/dsk/c6t0d2
/dev/dsk/c7t0d2
/dev/vg03
dKF1
/dev/dsk/c6t0d3
/dev/dsk/c7t0d3
/dev/vg04
dKF1
/dev/dsk/c6t0d4
/dev/dsk/c7t0d4
/dev/vg06
dKF1
/dev/dsk/c6t0d6
/dev/dsk/c7t0d6
/dev/vg01
dKF1
/dev/dsk/c6t0d1
/dev/dsk/c7t0d1
/dev/vg05
dKFB
/dev/dsk/c6t0d5
/dev/dsk/c7t0d5

The new disk names are below by running ioscan -funC disk

disk 0 0/0/2/0.0.0.0 sdisk CLAIMED DEVICE
/dev/dsk/c0t0d0 /dev/rdsk/c0t0d0
disk 1 0/1/1/0.0.0 sdisk CLAIMED DEVICE
3LC
/dev/dsk/c2t0d0 /dev/rdsk/c2t0d0
disk 2 0/1/1/0.1.0 sdisk CLAIMED DEVICE
3LC
/dev/dsk/c2t1d0 /dev/rdsk/c2t1d0
/dev/dsk/c2t1d0s1 /dev/rdsk/c2t1d0s1
/dev/dsk/c2t1d0s2 /dev/rdsk/c2t1d0s2
/dev/dsk/c2t1d0s3 /dev/rdsk/c2t1d0s3
disk 19 0/3/1/0.1.11.0.0.0.1 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d1 /dev/rdsk/c8t0d1
disk 22 0/3/1/0.1.11.0.0.0.2 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d2 /dev/rdsk/c8t0d2
disk 23 0/3/1/0.1.11.0.0.0.3 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d3 /dev/rdsk/c8t0d3
disk 24 0/3/1/0.1.11.0.0.0.4 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d4 /dev/rdsk/c8t0d4
disk 25 0/3/1/0.1.11.0.0.0.5 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d5 /dev/rdsk/c8t0d5
disk 26 0/3/1/0.1.11.0.0.0.6 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c8t0d6 /dev/rdsk/c8t0d6
disk 15 0/3/1/0.1.12.0.0.0.1 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d1 /dev/rdsk/c10t0d1
disk 16 0/3/1/0.1.12.0.0.0.2 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d2 /dev/rdsk/c10t0d2
disk 17 0/3/1/0.1.12.0.0.0.3 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d3 /dev/rdsk/c10t0d3
disk 18 0/3/1/0.1.12.0.0.0.4 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d4 /dev/rdsk/c10t0d4
disk 20 0/3/1/0.1.12.0.0.0.5 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d5 /dev/rdsk/c10t0d5
disk 21 0/3/1/0.1.12.0.0.0.6 sdisk CLAIMED DEVICE
SV110 (C)COMPAQ
/dev/dsk/c10t0d6 /dev/rdsk/c10t0d6

Is there a way to name the new disk names back to the original so I can mount the disks.

Thanks

[Marco A.]

Hello,

Actually I believe that they changed the way in which the disks were connected, is not about change the disk name, the name is determine depending of how are they connected.

Apparently you need to re-create the lvm structure.


Marco

[Mel Burslan]

Not to repeat what others already said about a similar problem, please see this thread:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=425790

yes, it is possible to change the hardware paths, but, if I were you, I would do one or maybe two ignite recovery tapes, as this procedure, done with a little imperfection, may render your system not bootable.

Hope this helps

[Senthil Kumar .A_1 · Accepted solution]

Hi,

We will follow the steps for /dev/vg02. if it goes as expected we will do it for the rest of the VG's

1:
vgexport -m /tmp/vg02_map /dev/vg02
2:
mkdir /dev/vg02
mknod /dev/vg02/group c 64 0x020000
3:
vgimport -v -m /tmp/vg02_map /dev/vg02 /dev/dsk/c8t0d2 /dev/dsk/c10t0d2
4:
vgchange -a y /dev/vg02
5:
vgdisplay -v /dev/vg02
6:
mount all the Lvols hosted from vg02

If the above steps are fine with vg02 then its time to repeat them for the rest of the VG's

SK

[Kaleung Tsou]

Hi Senthil,

I am getting this message on step 3

# vgimport -v -m /tmp/vg02_map /dev/vg02 /dev/dsk/c8t0d2 /dev/dsk/c10t0d2
Beginning the import process on Volume Group "/dev/vg02".
vgimport: Minor number of /dev/vg02/group is not unique. /dev/vg03/group has the same minor number.

Any suggestions?

Thanks

[A. Clay Stephenson]

By convention, vg01 uses 0x010000, vg02 uses 0x02000, ... vg10 uses 0x0a0000 but that is simply convention. Do an "ls -l /dev/vg*/group" to display all the current group monor device numbers and choose one that is not currently in use AND so that the first two hex digits are less than your maxvgs kernel tunable. Use that value as your monir device number and proceed with the vgimport.

[Kaleung Tsou]

Here is what I got from:

# ls -l /dev/vg*/group
crw-r----- 1 root sys 64 0x000000 Mar 21 09:40 /dev/vg00/group
crw-rw-r-- 1 root sys 64 0x060000 Apr 27 04:10 /dev/vg01/group
crw-rw-rw- 1 root sys 64 0x020000 May 9 10:04 /dev/vg02/group
crw-rw-r-- 1 root sys 64 0x020000 Apr 27 03:58 /dev/vg03/group
crw-rw-r-- 1 root sys 64 0x030000 Apr 27 04:06 /dev/vg04/group
crw-rw-r-- 1 root sys 64 0x070000 May 9 2007 /dev/vg05/group
crw-rw-r-- 1 root sys 64 0x040000 Apr 28 01:35 /dev/vg05old/group
crw-rw-r-- 1 root sys 64 0x050000 Apr 27 04:08 /dev/vg06/group
cr--r--r-- 1 root sys 64 0x040000 Apr 28 01:52 /dev/vg07old/group

Can I use 0x010000 for vg02 as the minor number? Sorry for being so ignorant, I am a newbie.

[A. Clay Stephenson]

Yes, rm /dev/vg02/group and then
mknod /dev/vg02/group c 64 0x010000

Note: That all except vg00 are really out of sync with convention so in the future you might consider vgexporting and vgimporting and reassigning the minor device numbers to avoid confusion.

[Kaleung Tsou]

I have all the vg back and I reassigned the minor numbers as suggested. However, when I try to mount below; I get a this error

# mount /dev/vg03/orabase /oracle
UX:vxfs mount: ERROR: V-3-21268: /dev/vg03/orabase is corrupted. needs checking

Thanks again.

[A. Clay Stephenson]

In that case, fsck -F vxfs /dev/vg03/rorabase

You may need to specify the -o full or -o full,nolog options. Man fsck_vxfs for details.

After fsck completes you should then be able to mount the file system.