LVM - mounting a disk against it's will

[Peetu]

I'm having trouble with a disk drive. Not sure what caused it, but the disk isn't anymore recognized and it's not getting mounted. I'm now trying to get the volume group 2 working again.

This far I've done
mv /etc/lvmtab /etc/lvmtab.old2008Nov
vgscan -a -v

The vgscan gives some nasty looking error.
Since I don't want to take time writnig it all I took a picture. It is attached.

Maybe I have to somehow update the list of drives? 'ioscan -funC disk' shows also an old tape drive (If I understand correctly)...

The 'ü' symbols are actually '-'... What's the cure for this problem, by the way :)
It doesn't matter much since I usually use SSH...

Here is a previous thread I wrote about some problems with the same drive. The problem was the same, it disappeared one day but I managed to get it working again.
http://forums11.itrc.hp.com/service/forums/questionanswer.do?admit=109447626+1225180209972+28353475&threadId=1220257

[Aneesh Mohan · Accepted solution]

Hi,

Can you post the output of

#vgcfgrestore -n vg02 -l;date /etc/lvmconf/vg00.conf

Aneesh

[Ninad_1]

Hi,

Can you also provide
strings /etc/lvmtab.old2008Nov
diskinfo /dev/rdsk/c3t12d0
ll /dev/rdsk/c3t12d0


Regards,
Ninad

[Matti_Kurkela]

According to the picture, you've run "ioscan -funC disk". It might show you old information, so try without the -u option: "ioscan -fnCdisk".

If that displays your disk /dev/dsk/c3t12d0 as NO_HW, then I'm afraid your disk has died. Does "diskinfo -v /dev/rdsk/c3t12d0" report anything useful?

If your disk is shown as NO_HW, the disk has probably failed its own self-tests and electronically disconnected itself from the SCSI bus. There are a few data recovery tricks you might try to get the disk started up one more time.

Try unplugging the disk from the SCSI tower and plugging it back in. Listen for the sounds the disk makes.

If the disk spins up, good. If not, you might try unplugging it and spinning the disk around in your hands, using the momentum to make the disk platters spin a little. If the disk's bearings are sticking, this might help to unstick them.

If you manage to get the disk come back alive, make a backup of it ASAP - it is obviously about to fail completely.

The /dev/dsk/c0t0d0 device is not a tape drive, it's a Teac CD-532E-B CD-ROM. It's probably internal to the workstation and equipped with a HP-branded front panel. Tape drives should never appear in any "ioscan -C disk" listing. To see the tape drive, run "ioscan -fnkC tape".

Regarding your problem with the "-" characters... what kind of terminal are you using? Its character set is probably mis-configured.

MK

[Dennis Handly]

>MK: what kind of terminal are you using? Its character set is probably mis-configured.

Probably some windows font? I sometimes see something like dominoes for "-". This could be something like short or long "-"?

[Peetu]

Here is the output of vgcfgrestore.

I'm sorry for the bad characters. The german y characters (u with a dot over it) are again supposed to be '-'.

[Peetu]

Here is the ioscan output.
It looks like removing the u didn't change anything.

[Peetu]

Here is the lvmtab, diskinfo and the permissions for the /dev/rdsk/c3t12d0.

The permissions read 'crw-r-----'.

The diskinfo with size=0Kbytes is a bit alarming.

Luckily, I have a backup of the files. I just don't have another SCSI drive ready so I'd like to get this drive working.

Do I have to go shopping soon? :)

[Aneesh Mohan]

Hi,

Try

#ioscan -fn

#insf -e

#mv /etc/lvmtab /etc/lvmtab.old

#vgscan -v

#strings /etc/lvmtab


Aneesh

[Steven E. Protter]

Shalom,

A quibble with the title.

LVM - mounting a disk against it's will

Disks are not sentient and have no free will. They do what your system configuration tells them to do.
Unless perhaps they are cylon (see battlestar galactica sites for reference).

What has happened here is your configuration is mounting a disk via LVM against your will, your intended configuration.

My guess had I caught this thread early would have been a configuration problem manifested in /etc/lvmtab

It would seem that you have a solution.

Good luck,

SEP

[Matti_Kurkela]

By my experience, if the diskinfo command reports the disk size as "0 Kbytes", that usually means the disk is dead.

When the disk's internal self-test fails, the disk will, depending on the exact disk model, either completely drop out of the SCSI bus or otherwise indicate a failure. The indication of "size: 0 Kbytes" would certainly be one way to indicate the disk is not usable any more.

MK

[Peetu]

Thanks to all of you so far!

Here is 'ioscan -fn'.

[Peetu]

Here is the insf output. The drive in question is highlighted.

[Peetu]

And finally, here is the result of moving the lvmtab and doing 'vgscan -v'.

There's no vg02 in the lvmtab... :)

I quess the drive has seen it's days?