VG00 does not exist in LVMTAB

[Quare]

Hi all,

i just want to know why when i ran vgscan -p -v the result as below:

vgscan: Couldn't access the list of physical volumes for volume group "/dev/vg00
".

when i see at SAM under disk and devices it stated that "unused". how can i can change the status to "VG00" since when i want to activate using vgchange the error comes stated that vg00 does not exist in lvmtab.

but when i perform bdf | more, i can see the vg00.

My question is:
1. How to make this vg00 exist in lvmtab
2. How to make the status in SAM is "vg00" in state of "unused".


Please help

[njia_1]

Hi there

Can not access physical volumes indicates a possible hardware problem.

Have you tried "string /etc/lvmtab" this should print out the vg configuration for vg00 and which diskes are included.

If the status of a disk is unused and you are sure about it is unused. You can add this disk to a vg by
Create a PV first on that disk
pvcreate /dev/rdsk/cxtxdx you may have to use pvcreate -f if that disk was used by another vg before.

then vgextend to extend VG to that disk.

thanks

[Rajeev Shukla · Accepted solution]

What it seems like is that lvmtab file has corrupt information. In which case you can try 2 things
1. Rename the /etc/lvmtab file and then do a vgscan
2. If your lvmtan file doesn't have information of vg00 which you can see by doing strings on /etc/lvmtab file then import the vg00 with the disks that you know are part of vg00
vgimport -v /dev/vg00

[Quare]

Hi njia,

why when i make bdf command i can see the vg00 is mount to file.Below is the output when issued bdf command:
/dev/vg00/lvol3 60% /
/dev/vg00/lvol8 24% /var
/dev/vg00/lvol7 75% /usr
/dev/vg00/lvol5 75% /home
/dev/vg00/lvol11 86% /oracle

i have issued strings /etc/lvmtab but there is no devices for vg00. i just don't understand why there is still file system mount on vg00 even the status is unused.

let say when i run pvcreate for vg00,is it data in file system that mount by vg00 is gone or still in there.

Please help

[Rajeev Shukla]

For sure your lvmtab file is corrupt and doesn't have information for VG00.
USe the above steps to build your lvmtab file

[Quare]

Hi rajeev,

i did run the vgscan and then vgimport. I still got message below:

Beginning the import process on Volume Group "/dev/vg00".
Volume group "/dev/vg00" is still active.
gst:/# vgchange -a n /dev/vg00
vgchange: Volume group "/dev/vg00" does not exist in the "/etc/lvmtab" file.

how to deactivate this vg00 since it cannot be seen at lvmtab. i try to unmount some file for vg00 but it stated device busy.

Any other suggestion on this?
Please help

[Rajeev Shukla]

you need to rename the /etc/lvmtab file to something else before you do a vgscan

[Quare]

yes, i did all what you stated but there still message that vg00 is still active.

in other way, do i have to unmount all the file system that point to vg00 and then vgimport and then activate it.

but one more thing occured when i try to unmount it the error device is busy. what else rajeev.

pls help

[rayche]

yes, some file system that you cannot umount, because vg00 is root vg! if you want to deactive it and if your machine can reboot, pls use hpux -lm to enter maitanence mode, in that mode, vg00 deactive.

Thanks.

[Adisuria Wangsadinata_1]

Hi Quare,

You cannot de-active vg00 if you're still in the multiuser mode. You need to boot the system with maintenance mode OR single user mode :

maintenance mode :
hpux -lm

single user mode :
hpux -is

BTW, check with ioscan command to see whether the boot disk with what status ?

# ioscan -nfCdisk

Hope this information can help you.

Cheers,
AW

[Quare]

Hi Rajeev,

in this case, if i use pvcreate does it affect anything to the system like data loss or other consequences.

since i cannot reboot the server, what else can i do to make it happen?

Regdrs

[Siju Vadakkan]

put the details.

vgdisplay -v vg00 | grep "PV Name"
strings /etc/lvmtab
ll /etc/lvmconf
strings vg00.conf
lvlnboot -v
ioscan
ll /etc/lvmta*

[Rajeev Shukla]

No dont do pvcreate. If you do so you will loose all the LVM information.
OK, after you have moved the /etc/lvmtab file to say /etc/lvmtab.ori and run vgscan what error or message do you get?
Do you see a lvmtab file created after you run vgscan..if yes do a strings and see the conents of the file.
And before you do so make sure you can see all the devices which are part of vg00 using ioscan -fnCdisk

[Quare]

Hi Siju,

1).gst:/# vgdisplay -v vg00 | grep /dev/dsk/c1t0d0
Information in "/etc/lvmtab" differs from that in "/etc/lvmpvg".
VG "/dev/vg00" is not defined in file "/etc/lvmtab".
vgdisplay: Volume group "/dev/vg00" does not exist in the "/etc/lvmtab" file.
vgdisplay: Cannot display volume group "vg00".

2)gst:/# strings /etc/lvmtab
/dev/vg01
/dev/dsk/c12t0d0
/dev/dsk/c12t0d1
/dev/dsk/c14t0d0
/dev/dsk/c14t0d1
/dev/vg02
/dev/dsk/c12t0d2
/dev/dsk/c12t0d3
/dev/dsk/c14t0d2
/dev/dsk/c14t0d3
/dev/vg03
/dev/dsk/c12t0d4
/dev/dsk/c12t0d5
/dev/dsk/c14t0d4
/dev/dsk/c14t0d5
/dev/vg04
/dev/dsk/c14t0d6
/dev/dsk/c14t0d7
/dev/dsk/c12t0d6
/dev/dsk/c12t0d7

3)gst:/# ll /etc/lvmconf
total 34048
---------- 1 root root 0 Jan 6 2004 lvm_lock
-rw------- 1 root root 321536 Apr 19 2004 vg00.conf
-rw------- 1 root root 321536 Apr 19 2004 vg00.conf.old
-rw------- 1 root sys 1725440 Mar 5 15:33 vg01.conf
-rw------- 1 root sys 1724416 Mar 5 15:00 vg01.conf.old
-rw-rw-rw- 1 root sys 132 Jan 15 2004 vg01.mapfile
-rw------- 1 root sys 1725440 Mar 5 15:38 vg02.conf
-rw------- 1 root sys 1724416 Mar 5 15:38 vg02.conf.old
-rw-rw-rw- 1 root sys 132 Jan 15 2004 vg02.mapfile
-rw------- 1 root sys 1725440 Mar 5 15:41 vg03.conf
-rw------- 1 root sys 1724416 Mar 5 15:40 vg03.conf.old
-rw-rw-rw- 1 root sys 132 Jan 15 2004 vg03.mapfile
-rw------- 1 root sys 1725440 Mar 5 15:44 vg04.conf
-rw------- 1 root sys 1724416 Mar 5 15:43 vg04.conf.old
-rw-rw-rw- 1 root sys 132 Jan 15 2004 vg04.mapfile
-rw------- 1 root sys 1723392 Mar 5 15:46 vg05.conf
-rw------- 1 root sys 1181696 Nov 4 2004 vg05.conf.old

4)gst:/# strings vg00.conf
vg00.conf: No such file or directory

5)gst:/# lvlnboot -v
Current path "/dev/dsk/c14t0d0" is an alternate link, skip.
Current path "/dev/dsk/c14t0d1" is an alternate link, skip.
Current path "/dev/dsk/c14t0d2" is an alternate link, skip.
Current path "/dev/dsk/c14t0d3" is an alternate link, skip.
Current path "/dev/dsk/c14t0d4" is an alternate link, skip.
Current path "/dev/dsk/c14t0d5" is an alternate link, skip.
Current path "/dev/dsk/c12t0d6" is an alternate link, skip.
Current path "/dev/dsk/c12t0d7" is an alternate link, skip.
Current path "/dev/dsk/c12t1d0" is an alternate link, skip.

[rayche]

1.Could u use dd command to test the error disk first ?

dd if=/dev/dsk/cxtxdx of=/dev/null bs=1024k

2. to your question, first, I think your lvmtab corrupted, second, in fact, if your system run normally, you can move your lvmtab, this do not affect machine when the machine is running, so although your lvmtab corrupted, lvs of vg00 run normally

[Quare]

Hi Rajeev,

Attach is the file of vgscan.

there is nothing created on file /etc/lvmtab.

for the ioscan -knfCdisk, yes i can see all the claimed device.

Rgdrs

[Rajeev Shukla]

Yes there is another file /etc/lvmpvg which i beleive has wrong information and its wise to rename that as well before you do a vgscan.
So try moving both lvmtab and lvmpvg files to different names and do a vgscan. I am sure it should do the magic

[Quare]

Hi rajeev,

Thanks rajeev, it's really make a magic.
Now i can see at SAM that it stated VG00 and also i can see VG00 at /etc/lvmtab.

Just wanna ask you if you have an idea,why could this happen?

Is it possible it happens again if i reboot the server.

Last to ask, what is the best to do after we activate the volume group.

Thanks...

[Rajeev Shukla]

Glad to hear that it solved your problem but i too am not sure as why it happens, but its true that its some sort of corruption with the lvmtab file which conflicts with the running kernel and lvmpvg file. Now to Guarantee that it doesn't happen again, not sure..but yes now that it is fixed, a reboot wont cause the problem again.

Dont forget to assign points to people who put their efforts in.

Regards
Rajeev

[Quare]

Hi rajeev,


ok thanks..

[Quare]

The problem has been resolved thru the solution given by the members in the forum