Lost VG(s)

[Olga_1]

I have installed QP11.00 on HP 11.00 and after the reboot some of the volume groups were not mounted. I looked at SAM and I have three disks unused.
Is there any way to restore the configuration and to find out what happend? Also after reboot I received only one EMS message, usualy I have 6-7 of them.

Thank you.

[Pete Randall]

Olga,

Try running vgscan. First rename /etc/lvmtab "mv /etc/lvmtab /etc/lvmtab.old", then run "vgscan -a" and see what you get.


Pete

[Massimo Bianchi]

Hi, before doing abything else, try activating manually the vg.

Maybe someone has changed the /etc/lvmrc from

AUTOVG_ACTIVATE from 1 to 0.


Check also that !

If that does not help, follow SEP suggestions..

HTH,
Massimo

[Pete Randall]

Massimo,

That's twice now. I'm not SEP! SEP is Steven Protter. I'm Pete Randall.

;^)


Pete

[PIYUSH D. PATEL]

Move the /etc/lvmtab file and run vgscan -a.

HTH,
Piyush

[Olga_1]

Here is theoutput of vgscan -a:

Creating "/etc/lvmtab".
Couldn't stat physical volume "/dev/dsk/c3t2d0":
Invalid argument
Physical Volume "/dev/dsk/c4t0d0" contains no LVM information
Physical Volume "/dev/dsk/c5t0d0" contains no LVM information


Following Physical Volumes belong to one Volume Group.
Unable to match these Physical Volumes to a Volume Group.
Use the vgimport command to complete the process.
/dev/dsk/c2t2d0
/dev/dsk/c4t1d0
/dev/dsk/c5t1d0


Following Physical Volumes belong to one Volume Group.
Unable to match these Physical Volumes to a Volume Group.
Use the vgimport command to complete the process.
/dev/dsk/c4t9d0
/dev/dsk/c5t9d0


*** LVMTAB has been created successfully.
*** If PV links are configured in the system.
*** Do the following to resync information on disk.
*** #1. vgchange -a y
*** #2. lvlnboot -R

[Massimo Bianchi]

Gee.... I truly apologize !!!

I know the difference between you two, but I made a very fast read of the answer, and i was thinking of some other possibility.

When writing, i only remember that there "e" in the name and wrote SEP...

PETE !!!! Other public apologies..
Massimo

[Pete Randall]

It would seem that you now need to run vgimport on /dev/dsk/c4t9d0 and /dev/dsk/c5t9d0.


Pete

[Massimo Bianchi]

Hi,
i'm a little worried of this behaviour.
Even if you are able to rebuild your lvm config, following MIGHTY PETE hints [I owe him some :) ], i'm asking what has happened.

Is there anything strange in the syslog, like "device found" or "installing special files for..." entry ?

HTH,
Massimo

[PIYUSH D. PATEL]

Hi,

Do you have any information regarding which disks belong to which volume group ?

If yes, the you can use the vgimport command.

If not then you need to create the vg first and then vgimport the disks.

HTH,
Piyush

[Pete Randall]

Massimo,

No problem - don't worry about it!


Olga,

Any errors during startup? Check dmesg, /etc/rc.log, /var/adm/syslog/syslog.log.


Pete

[Bernhard Mueller]

Olga,

first I would not worry about EMS taking a while to send the restart mail.

However, if it took considerable longer than usual, you might have a HW problem somewhere.

Can you verify that you see all the disks you have in ioscan (plus all alternate links)?
(ioscan -fnCdisk)?

If you have them all, it may be a SAM problem showing alternate paths as "unused"

Do you have a documentation about which disks belong to which VG (strings /etc/lvmtab)?

Do you have all the device files of the (not activated) VGs (ll /dev/vg*/group; ll /dev/vg*/*)

Supposing everything looks OK just try to vgchange -a y vgXY, and issue mountall to see whether the LVM information on the disks is OK.

If, however, SAM is right, then you may have to use vgcfgrestore to write the LVM headers from /etc/lvmconf/vg*.conf back to the disks (if you know which belongs to which).

Regards,
Bernhard

[James R. Ferguson]

Hi Olga:

It appears that your volume group device files may have disappeared. Do:

'ls -l /dev/vg*'

...to see if you are missing the volume group that you had before. If so, recreate the device file and 'vgimport'. I presume that you have paper documentation of your server's configuration. If not, look at your last Ignite manifest:

/var/opt/ignite/recovery/latest/manifest

The above is generated automatically every time you perform an Ignite 'make_tape_recovery'.

Regards!

...JRF...

[Bernhard Mueller]

Oh,

it took me so long to write the message...
I need to catch up with what has happened.

So you have a new /etc/lvmtab, hope you kept the old one around, pls do
# strings /etc/lvmtab.old
(or however you called it)

# ll /dev/vg*/group

# cat /etc/fstab

and let us know the results.

(I take it this is NOT a Cluster Node)

Regards,
Bernhard

[Olga_1]

Here is some information that you asked:

ll /dev/vg*/group
two VG are missing here vg02 and vg04

crw-r----- 1 root sys 64 0x000000 Oct 22 2001 /dev/vg00/group
crw-r--r-- 1 root sys 64 0x030000 Feb 7 2002 /dev/vg03/group
crw-r--r-- 1 root sys 64 0x050000 Oct 3 2002 /dev/vg05/group


strings /etc/lvmtab.old
/dev/vg00
/dev/dsk/c1t2d0
/dev/dsk/c1t0d0
/dev/dsk/c2t0d0
/dev/vg01
/dev/dsk/c5t0d0
/dev/vg02
/dev/dsk/c5t1d0
/dev/dsk/c2t2d0
/dev/vg04
/dev/dsk/c5t9d0
/dev/vg03
/dev/dsk/c4t8d0
/dev/dsk/c5t8d0
/dev/dsk/c4t2d0
/dev/dsk/c5t2d0
/dev/vg05
/dev/dsk/c4t10d0
/dev/dsk/c5t10d0

Here I have the location of the VG02 and VG04

So based on this can I run vgimport?

If I do /usr/sbin/vgimport -v -m /etc/lvmconf/vg04.mapfile /dev/vg04
Beginning the import process on Volume Group "/dev/vg04".
vgimport: "/dev/vg04/group": not a character device.

[Massimo Bianchi]

Hi,
you must recreate the special files for

vg02 and vg04

mknod /dev/vg02/group c 64 0xNN0000
mknod /dev/vg04/group c 64 0xNN0000


where NN are your free minor numbers.
and then repeat the vgimport

HTH,
Massimo

[Pete Randall]

Olga,

First you need to rebuild the device files thusly:

mkdir /dev/vg04
mknot /dev/vg04/group c 64 0x040000

(repeat for the other missing VGs)

Then run the vgimport.


Pete

[Olga_1]

If I do
vgcfgrestore -n /dev/vg04 /dev/rdsk/c5t9d0

c5t9d0 I got from strings /etc/lvmtab.old

Will it work?

[Olga_1]

If I do
vgcfgrestore -n /dev/vg04 /dev/rdsk/c5t9d0

c5t9d0 I got from strings /etc/lvmtab.old

Will it work?

[Dagmar Boelen]

Hi,

First recreate the volumegroup-file. You'll need them!

[PIYUSH D. PATEL]

Hi,

After creating the VG04 you can do :

vgcfgrestore -n /dev/vg04 /dev/rdsk/c5t9d0
vgchange -a y /dev/vg04

But since the vg04 information is not there, vgcfgrestore may/may not work.

Use the vgimport command instead.

vgimport /dev/vg04 /dev/rdsk/c5t9d0
vgchange -a y /dev/vg04

Then you can mount the filesystems.

HTH,
Piyush

[Olga_1]

Here is what I did:

mknod /dev/vg04/group c 64 0x040000

mknod /dev/vg02/group c 64 0x020000

ll /dev/vg*/group
crw-r----- 1 root sys 64 0x000000 Oct 22 2001 /dev/vg00/group
crw-r--r-- 1 root sys 64 0x020000 Jul 14 10:57 /dev/vg02/group
crw-r--r-- 1 root sys 64 0x030000 Feb 7 2002 /dev/vg03/group
crw-r--r-- 1 root sys 64 0x040000 Jul 14 10:55 /dev/vg04/group
crw-r--r-- 1 root sys 64 0x050000 Oct 3 2002 /dev/vg05/group

/usr/sbin/vgimport -v -m /etc/lvmconf/vg04.mapfile /dev/vg04
Beginning the import process on Volume Group "/dev/vg04".
vgimport: Unable to read the physical volume

Any ideas?

[Massimo Bianchi · Accepted solution]

Hi,
you must specify the list of disks.


vgimport -v -m /..... vg02 /dev/dsk/... /dev/dsk/....


HTH,
Masimo

[James R. Ferguson]

Hi Olga:

You need to specify the device files during the 'vgimport':

# /usr/sbin/vgimport -v -m /etc/lvmconf/vg04.mapfile /dev/vg04 /dev/dsk/c5t9d0

...based on your post of your old '/etc/lvmtab'.

Regards!

...JRF...

[Dagmar Boelen]

Hi,

I think there is still a problem with the disks. The LVM-info on these PV's are not correct. Use the vgcfgrestore.