external mount of a logical volume

cygnus-x1 · 07-25-2006, 01:01 PM

Hey guys. I am new to the LVM that Fedora automatically uses when it installs itself. I am trying to look at a hard disk that is not coming up. I booted off of a SlackWare install CD and I was able to mount the /dev/hda1 (boot partition) but I could not mount the /dev/hda2 (root) partition. When I did fdisk I saw that /dev/hda2 was setup as Linux LVM type. The error I saw on the screen when he was trying to boot was a message about ext3 so i assume that the underlying fs is ext3. So I tried mounting it like this:

mount -t ext3 /dev/hda2 /mnt/steve2

That did not work so I looked into the grub.conf file and found the LVM format

mount -t ext3 /dev/VolGroup00/LogVol00 /mnt/steve2 (I know grasping at straws)

Can anyone help me mount this puppy so I can look at it?
If I get it mounted what can I do to check it out? Will fsck work on an LVM? Any and all help appreciated.

thanks

Doug

mjgsantos · 07-25-2006, 03:03 PM

Hi,
This is a very boring procedure, but may work:
When you use the logical volume manager, it creates a Volume Group that can have one or more hard disks assigned. Then, it creates a Physical Volume, that may be alocated on one or more disks, and then is created a Logical Volume, that is where the file system will be created a mounted from.
There is one problem: the Linux OS need to know the volume group configuration to be able to mount the filesystem from the /dev/VolGroup00/LogVol00. Try to execute the command vgscan, so it will look for any volume group in the current disks:

Code:

% vgscan
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- found active volume group "notes5prd"
vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
vgscan -- WARNING: This program does not do a VGDA backup of your volume group

Then, try to execute the command pvdisplay passing the device where the LVM was created (i.e: pvdisplay /dev/hda2). You may run this command in each device displayed on fdisk -l command. This command will show the vg namem if there is any information on the device that you chose.
Finally, run the command vgimport followed by the device(s) that contains any vg information. This will enable the OS to recognize the LVM configuration. If this works, you will be able to execute the vgdisplay command and see the vg status. You will need to run vgchange -a y to enable the volume group, and then you be able to mount the filesystem from this crazy mount point.
AFAIK, fsck works the same way.
Please tell me later if this worked.

Regards.

cygnus-x1 · 07-26-2006, 03:28 PM

I was able to do a vgscan but I could not do a vgdisplay.
The output of fdisk -l had odd looking device names (not /dev/hda etc)
I still tried using those but was still unable to run it.

When I reboot the kernel loads and the it tries to load the root partition. When it does that I get a bad superblock error. Is this disk bad or can it be wiped and reused?

thanks

haertig · 07-26-2006, 03:51 PM

I assume Fedora uses LVM2 (just a guess). Does Slackware have LVM support by default? Is it LVM2?

Assuming you have LVM2 support on your Slackware LiveCD, boot it and try running:

Code:

# pvscan
# vgchange -a y
# pvs
# vgs
# lvs

And post the output here if you need more help. Also, "fdisk -l" output as well.

I suspect the "vgchange -a y" command is your missing link. You need to flag the volume group as "active". Your Slackware LiveCD probably does not know it needs to do this. Your Fedora installation probably does it in one of it's startup scripts.

Quote:

mount -t ext3 /dev/VolGroup00/LogVol00 /mnt/steve2 (I know grasping at straws)

Nothing wrong with this if you've gotten the volume group and logical volume names correct.

cygnus-x1 · 07-26-2006, 08:38 PM

Actually the Slackware install CD did not have LVM support built into the kernel. So I used a copy of SMART Linux that we had laying around. That did have LVM support but I was still unable to mount anything. I will try your suggestions tomorrow.

But my question still stands ... When you see a messag like bad superblock when you boot does that usually mean a bad disk or a corrupt filesystem?

thanks Doug

haertig · 07-27-2006, 10:56 AM

Quote:

But my question still stands ... When you see a messag like bad superblock when you boot does that usually mean a bad disk or a corrupt filesystem?

This is Fedora giving you this error, not your Slackware LiveCD, right? Just so we're clear.

It could simply mean, e.g., that you're trying to mount an ext3 filesystem but telling mount that it's reiser. This scenerio would be terribly unlikely if it's Fedora's boot sequence that's giving this error (it's fstab would surely know your filesystem types). But not so unlikely if you're trying to manually mount the thing via Slackware and are guessing at a filesystem type.

Filesystem corruption could cause this error too. This corruption may or may not have been caused by a bad disk. You can't say just based on the bad superblock error message. What happened around your place immediately prior to this problem occurring? A power hit? Totally unexpected system crash? Once you get something mounted and can look at your syslog, does it mention anything about read/write/seek disk errors?

Most (?) disk manufacturers have a non-destructive diagnostic utility available on their websites. You can use this to check out your disk. Seagate has one named "SeaTools". You create a bootable floppy or CD with this tool on it and run the test.

BTW, you have alternate superblocks available on your filesystems. If the primary one got corrupted, you might be able to recover from an alternate. Check out the fsck manpage for how to access alternates.