/dev/root

[Darren Murray_1]

I have a hp workstation system which is mounting / as /dev/root instead of /dev/vgroot/root.

I have tried copying /etc/mnttab to /etc/mnttab.orig and then doing a mount -a and bdf but it still shows as /dev/root.

It is still showing this as /dev/root even when not in maintenance mode.

Any ideas?

[Sunil Sharma_1]

what version of HP UX is running there ?

[Darren Murray_1]

its running 10.20

[Robert-Jan Goossens]

Hi Darren,

If it is possible, reboot your workstation.

Regards,
Robert-Jan

[Darren Murray_1]

have done that already..still the same

any other ideas

[Sridhar Bhaskarla]

Hi Darren,

If there is a problem with the disk associated with / filesystem, /dev/root gets mounted. It could also happen if there are any incorrect or corrupted entries in /etc/fstab. Try the commands 'lvlnboot -v', 'vgdisplay -v', 'pvdisplay -v ' 'lvdisplay -v ' etc., to see if there is anything abnormal.

-Sri

[Darren Murray_1]

Have checked the /etc/fstab file and its correct. All vgdisplay, lvdisplay , pvdisplay outputs are correct.

[Robert-Jan Goossens]

Could you post your /etc/fstab ?

[Ivajlo Yanakiev]

It can be all disk partition

send as /etc/fstab
also
vgdisplay -v

try and sent as :
lssf /dev/root

[Sridhar Bhaskarla]

Any errors on 'mount -a' command?. Any errors on the console during the boot time?. In /etc/rc.log?. Any 'stale' entries in your 'lvdisplay' or 'vgdisplay'?.

-Sri

[Darren Murray_1]

No errors occur when the server boots.
The lvdisplay shows no stale extents.
Checked rc.log and there are no errors in there either.

If I boot the machine into single user mode i get sh errors when trying to execute any command. If I boot the machine into maintenance mode and then do a vgchange -a y /dev/vgroot and do a mount -a I can then execute commands.

[Sridhar Bhaskarla]

Hmmm 'sh errors'?. Can you post the exact messages you are getting?. If you booted the system first into maintenance mode and then into multi-user, then you cannot get rid of /dev/root until you reboot it normally.

-Sri

[Darren Murray_1]

the fstab looks like this

/dev/vgroot/root / hfs defaults 0 1
/dev/vgroot/usr /usr hfs defaults 0 2
/dev/vgroot/opt /opt hfs defaults 0 2
/dev/vgroot/home /home hfs defaults 0 2
/dev/vgroot/var /var hfs defaults 0 2

[Sridhar Bhaskarla]

Hi,

Don't you have a /stand filesystem?. What do you have your 'boot' volume defined as?. 'lvlnboot -v'.

I see the non-default names for the vg00 LVs like (root, usr etc.,). I actually like it and configure them on my systems too. But during the reconfiguration there is a good chance that we may miss to rename some lvols and that can make /etc/fstab to go stale.

?? any clues there?

-Sri

[Darren Murray_1]

There is no /stand file system but thats standard for the workstation and its age Im pretty sure.

[Robert-Jan Goossens]

Darren,

Take a look at this doc.

http://www5.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000073839205

Document description: Migrating Root Volume Group to new disk(s).
Document id: LVMKBRC00007285

Regards,
Robert-Jan

[Darren Murray_1]

I dont have a spare disk unfortunately to do this to. Any other ideas??

[Sridhar Bhaskarla]

Darren,

What are those 'sh errors' you are getting when you logon using single-user mode?.

-Sri

[Darren Murray_1]

The sh errors are like

sh: more: not found
sh: bdf: not found

[John Palmer · Accepted solution]

Your sh errors are caused because /usr isnt't mounted. They're nothing to do with your original problem.

How long has the server been in this state? It's obviously not a standard installation because you've got vgroot instead of vg00.

The root filesystem doesn't actually get mounted because of its entry in /etc/fstab, it's actually located using pointers held in the volume group (reported by lvlnboot -v).

I suspect that there may be something amiss with the way /dev/vgroot/root is defined.

Could you please post the output from the following:
- lvlnboot -v
- ll /dev/vgroot
- vgdisplay -v vgroot

Regards,
John

[Ivajlo Yanakiev]

Use Make_recovery
create tape and reinstall your system using this archive. It will let you recovery all mess

[Darren Murray_1]

Do you know why /usr wouldnt be mounting??

If I go into maintenance mode then I can I can mount all file systems and run all commands fine. I then run a init 3 and then I cant run the commands. Would /usr be remounted on the init 3 and fail??


Will I be able to run successfully a make_tape_recovery from maintenance mode as no other level works. Also when I restore wont I just be restoring the same problem??

[Ivajlo Yanakiev]

NO you must make recovery from multi user mode.
Make_recovery make you bootable tape that you can use for installing OS and corect all problem.
You must stop auto recovery and you can make install not full recovery.
This mean that you can change any things.
All user data that are on root VG will be safe.
Use -A option for all VG
Read make_recovery man page.

[Darren Murray_1]

The problem is I cant get into multi user mode. When I try thats when I cannot run any commands.

[Ivajlo Yanakiev]

sent
#mount
command
and
screen that you see on standart boot