vg00 no more in /etc/lvmtab

[Mark Nieuwboer]

Hi,

I have a weird problem.
I used a vgdisplay -v vg00 and i get the messages:
vgdisplay: Volume group "/dev/vg00" does not exist in the "/etc/lvmtab" file.
vgdisplay: Cannot display volume group "vg00".

after this a moved /etc/lvmtab /etc/lvmtab.old
and did a vgscan -av know i had an other fault:
vgscan: has no correspoding valid raw device file under /dev/rdsk.
Verification of unique LVM disk id on each disk in the volume group
/dev/vg00 failed.

output ioscan -fnCdisk

disk 0 0/0/0/3/0.5.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t5d0 /dev/rdsk/c0t5d0
disk 1 0/0/0/3/0.6.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t6d0 /dev/rdsk/c0t6d0

/dev/rdsk# ll
total 0
crw-r----- 1 bin sys 188 0x005000 May 9 07:44 c0t5d0
crw-r----- 1 bin sys 188 0x006000 May 9 07:44 c0t6d0

dev/dsk# ll
total 0
brw-r----- 1 bin sys 31 0x005000 Nov 12 2004 c0t5d0
brw-r----- 1 bin sys 31 0x006000 Nov 12 2004 c0t6d0

All my filesystems are up on vg00

please how can i get vg00 back in /etc/lvmtab.

grtz. Mark

[Bharat Katkar]

Hi Mark,
Looks like one of the disk gone bad.

Well what you can do is just move all backup lvmtab files away from /etc and put them in /tmp.

Then try doing vgscan -a to recreate new lvmtab file.

If it still gives error then find which disk have gone bad using dd command.

# dd if= of=/dev/null

See if this tracks any bad block on the disk.

Try this for disk1 also.
Hope that helps.
Regards,

[Cem Tugrul]

Hi Mark,

Could you please check;

i-ll /dev/vg*

ii-cat /etc/fstab

iii-cat /etc/mnttab

iv-/etc/rc.log ?

[Sudeesh]

Hi,

You can try the following...

insf -eC disk
#to recreate device files
vgscan -av
#try creating the lvmtab.

If this fails verify that you are able to access both the hard disks using dd like Bharat suggested.



Sudeesh

[Saravanan_15]

Hi Nieu,
you can restore the vg00 configuration using vgcfgrestore command. First u can list the PVs that are belongs to vg00 from the config(backup)file.
For Ex:
Beta3dev /etc =>vgcfgrestore -n vg00 -l
Volume Group Configuration information in "/etc/lvmconf/vg00.conf"
VG Name /dev/vg00
---- Physical volumes : 3 ----
/dev/rdsk/c2t0d0s2 (Bootable)
/dev/rdsk/c2t1d0 (Non-bootable)
/dev/rdsk/c3t2d0 (Non-bootable)

By default it will take the inf. from the file /etc/lvmconf/vg00.conf

If you want to specify some other file,
Beta3dev /etc =>vgcfgrestore -f /etc/lvmconf/vg00conf.3103 -l
Volume Group Configuration information in "/etc/lvmconf/vg00conf.3103"
VG Name /dev/vg00
---- Physical volumes : 2 ----
/dev/rdsk/c2t0d0s2 (Bootable)
/dev/rdsk/c2t1d0 (Non-bootable)

Once you confirmed that file having enough inf. about your vg00, u can restore vg00 using,
Beta3dev /etc =>vgcfgrestore -R -n /dev/vg00 /dev/rdsk/c2t1d0
Volume Group configuration has been restored to /dev/rdsk/c2t1d0

Note: -R option will restore the config even if the VG is still in active.
Hope this will help you

[Cem Tugrul]

agree with Sudeesh's steps...

[Mark Nieuwboer]

Maybe i wasn't clear.
I already tried a vgscan with al the options.
I tried the dd options but no fault messages.
the vgcfrestore give the following output:
vgcfgrestore: Volume group "/dev/vg00" does not exist in the "/etc/lvmtab" file.
Volume Group configuration has been restored to /dev/rdsk/c0t5d0

the disk are fine but still i can't get vg00 in the lvmtab.

[Bharat Katkar]

Hi,

Something wrong with "c0t6d0"

Now that you have executed vgcfgrestore does command "vgdisplay -v vg00" works and if so can you post the output.

Also try executing "pvdisplay -v /dev/dsk/c0t6d0" and post the output.

Regards,

[Mark Nieuwboer]

I have also tryed to restore c0t6d0 same result.
vgdisplay -v vg00 gives still the same result.

something extra it is a cluster configuration.

[Mark Nieuwboer]

the output from pvdisplay =
pvdisplay -v /dev/dsk/c0t6d0
pvdisplay: Couldn't find the volume group to which
physical volume "/dev/dsk/c0t6d0" belongs.
pvdisplay: Cannot display physical volume "/dev/dsk/c0t6d0".

[Devender Khatana]

Hi,

You just mentioned that it is a cluster configuration. Is it a cluster configuration you are trying to set up? If yes on first node if you have done vgexport without -p option then it will remove information from lvmtab. I doubt that you are not doing this.

Just clarify your situation a bit. Are packages running ?

HTH,
Devender

[Saravanan_15]

Nieu,
You can try to import the vg. Try this one,

# vgimport -v /dev/vg00 /dev/dsk/c0t5d0 /dev/dsk/c0t6d0


good luck...

Regards,
Saravanan M

[Mark Nieuwboer]

The cluster is running fine also the packages.
I didn't do a vgexport of vg00 luckely.
The cluster is a week old.
a read something about
crw-rw---- 1 root root 64 0x000000 May 23 12:52 slvmvg
if this is wrong what to do.

also tried a vgimport of vg00 but what i thought is right you can't vgimport an actief vg.

[Devender Khatana]

Hi,

Oh yes, I think vg00 should be a cluster VG if your system is booting through lvm. It is a misconfigured system. Have you tested package switching etc. before finally handling it over.

It actually require a bit more information to tell everything in detail.

Attach the following :-

$strings /etc/lvmtab
$ioscan -fnC disk
$cmviewcl -v
On which node is your package running. Is it on the same node where problem exists?
How many nodes form this cluster?

HTH,
Devender

[Devender Khatana]

My applogies.
I mean to say vg00 should not be a cluster VG aas system boots from it and therefore other node can not activate it in exclusive mode.

HTH,
Devender

[Mark Nieuwboer]

Hi,

vg00 is not a cluster volumegroup.
for the cluster we have separte vg's.
the problem is not an the second cluster.
All the other vg the systems sees in the lvmtab except vg00.
it's a 2 node cluster with 4 packages equal shared over the cluster.
vg00 Is not a volumegroup of a clusterpackages.

grtz. Mark

[Senthil Kumar .A_1]

output ioscan -fnCdisk

disk 0 0/0/0/3/0.5.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t5d0 /dev/rdsk/c0t5d0
disk 1 0/0/0/3/0.6.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t6d0 /dev/rdsk/c0t6d0

the output from pvdisplay =
pvdisplay -v /dev/dsk/c0t6d0
pvdisplay: Couldn't find the volume group to which
physical volume "/dev/dsk/c0t6d0" belongs.
pvdisplay: Cannot display physical volume "/dev/dsk/c0t6d0".

Hi nieu,

These were your previous responses. The system has two HDD and the second output shows that, the c0t6d0 is not a part of LVM. So lets assume that c0t5d0 is a part of LVM and surely sould be part of VG00.

I need the following details.

1) ll -l /dev/vg*
2) strings /etc/lvmtab ( current lvmtab file ,even if it is corrupt).
3) lssf /dev/rdsk/c0t6d0
4) lssf /dev/rdsk/c0t5d0
5) pvdisplay -v /dev/dsk/c0t5d0

These infos i hope might suffice to check whats the problem...
Will wait for ur response.

[Cem Tugrul]

As an addition to Senthil,

You may also need to check;
#lvlnboot -v

-v Print verbose messages. With no other
arguments present, print information on root,
boot, swap, and dump logical volumes. If a
combined root-boot volume is configured, no
information for the boot volume is displayed.

[Ermin Borovac · Accepted solution]

I think this is a known problem. Are you running 11.00?

If so, check if you have the following patch installed?

Patch Name: PHCO_24437
Patch Description: s700_800 11.00 LVM commands cumulative patch

*** START PATCH NOTES ***
( SR: 8606111918 CR: JAGab84052 )
The vgscan command puts a bogus entry in the lvmtab
file, when a special file /dev/slvmvg is present
with the same minor number as /dev/vg00.

( SR: 8606111918 CR: JAGab84052 )
vgscan looks for all special files with major number
0x64 under /dev and assumes that the directory in which
it is present is a volume group name.

Resolution:
The vgscan command is modified to check the group file
under each volume group directory.
*** END PATCH NOTES ***

/etc/slvmvg is created by SG. You can move it away to /tmp (stop the cluster) and re-run vgscan. slvmvg should be re-created when you restart cluster with cmruncl.

[Saravanan_15]

Hi Nieu,
From the error messages reported by vgscan command,
"vgscan: has no correspoding valid raw device file under /dev/rdsk.
Verification of unique LVM disk id on each disk in the volume group
/dev/vg00 failed."

it seems to be problem with raw disk device files. and also see the timestamp of your raw device files,

crw-r----- 1 bin sys 188 0x005000 May 9 07:44 c0t5d0
crw-r----- 1 bin sys 188 0x006000 May 9 07:44 c0t6d0

It's showing may 9. Any modification you did.try to remove the raw device files using 'rmsf' command, and re-install the special files using "insf -e". Now move the lvmtab file, and run "vgscan -av".

Regards,
Saravanan M

[Henk Geurts]

hi Mark
running HPUX 11.11?
are you patch levels oke?
i had a similar problem once.
problem was that /etc/slvmvg had the same minor number as vg00, with an earlier time-stamp, therefor at boot it cannot find the vg. (thanx to installing MC/SG)
Ermin is right. : stop the cluster, remove /etc/slvmvg. and recreate /etc/lvmtab.

regards.
Henk

[Mark Nieuwboer]

we had to install the server from scratch