Operating System - HP-UX
1836380 Members
2505 Online
110100 Solutions
New Discussion

lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

 
Jack C. Mahaffey
Super Advisor

lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

Here's my dilemma.
We have two boot paths. 1st boot path boots to 10.20 with a mirrored disk. 2nd boot path boots to 1 disk. Older disk is 10.20. Newer is 11.0.
We are currently interrupting the boot sequence to come up on second disk. (11.0). I'm in the process of recovering the physical volume from the older boot disk. I've done the following:
1-Stop the mirroring on the old disks
lvreduce -m 0 /dev/vg00/lvol1
...
lvreduce -m 0 /dev/vg00/lvol1
2-Detach the old physical volume from /dev/vg00
vgreduce /dev/vg00 /dev/dsk/c1t5d0
3-Add the physical volume to the 11.0 boot
vgextend /dev/vgroot /dev/dsk/c1t5d0
4-Start mirroring the 11.0 system directories
lvextend -m 1 /dev/vgroot/boot /dev/dsk/c1t5d0
...
lvextend -m 1 /dev/vgroot/lvol9 /dev/dsk/c1t5d0
5-Ran lvnboot -v to check boot info and got
Root: ???
No Swap volume configured.
No Dump Volume configured.

What did I do to mess up the lvlnboot configuration?
I did notice that lvextend responded with Warning: Logical Volume Number "1073741827" found on physical volume not found in "/dev/vgroot".
This is a production server. The additional boot disk now shows up in lvlnboot. I'm now afraid the server will not boot.
Should I do lvreduce -m -0 for each mirrored logical volume, run vgreduce /dev/vgroot /dev/dsk/c1t5d0 and then start over with a pvcreate -f option? I did not use the -f when I ran vgreduce against the old boot disk.
Any help will be deeply appreciated. What makes matters worse is I don't have an attached tape drive for make_recovery tapes.
5 REPLIES 5
W.L. Alsemgeest
Occasional Advisor

Re: lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

This is what You should have done:
pvcreate -B (new disk)
vgextend vg00 (new disk)
lvextend -m 1 /dev/vg00/lvol# (for all the lvols)
mkboot /dev/rdsk/special_file
mkboot -a "hpux -lq" /dev/rdsk/special_file
lvlnboot -b /dev/vg00/(your boot lvol##)
lvlnboot -r /dev/vg00/(your root lvol##)
lvlnboot -s /dev/vg00/(your swap lvol##)
lvlnboot -d /dev/vg00/(your dump lvol##)
lvlnboot -R /dev/vg00

Greetings Wim Alsemgeest
Andy Monks
Honored Contributor

Re: lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

From where you are now, just doing from the mkboot of Wim's instructions should work. So :-

mkboot /dev/rdsk/special_file
mkboot -a "hpux -lq" /dev/rdsk/special_file
lvlnboot -b /dev/vg00/(your boot lvol##)
lvlnboot -r /dev/vg00/(your root lvol##)
lvlnboot -s /dev/vg00/(your swap lvol##)
lvlnboot -d /dev/vg00/(your dump lvol##)
lvlnboot -R /dev/vg00

You might want to start with 'lvrmboot -r' first so to be in a known good point.
Sandor Horvath_2
Valued Contributor

Re: lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

Hi !

Do next in 3. step

pvcreate -f -B /dev/rdsk/.....
vgextend vg00 /dev/dsk/....
mkboot /dev/rdsk/....
mkboot -a "hpux -lq" /dev/rdsk/...

regards, Saa
If no problem, don't fixed it.
Jack C. Mahaffey
Super Advisor

Re: lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

Thanks to all.
Jack C. Mahaffey
Super Advisor

Re: lvnboot -v reports ??? for root and No Swap/Dump after using vgextend

I went back and did the following in order to use the -B option for pvcreate.

# lvreduce -m 0 /dev/vgroot/lvol...
# vgreduce /dev/vgroot /dev/dsk/c1t5d0

I then did the following:

# pvcreate -fB /dev/rdsk/c1t5d0
# vgextend /dev/vgroot /dev/dsk/c1t5d0

I'm still got the message "Warning: Logical Volume number "107341827" found on physical volume not found in "/dev/vgroot".

Is this telling me that pvcreate did not properly prepare the disk? Should I be concerned?

I also have questions on the remaining steps, since the remaining steps are performed after vgextend has been used on the active production volume group.

My next steps are as follows:
# lvextend -m 1 /dev/vgroot/lvol# (for all logical vols)
mkboot /dev/rdsk/c1t5d0
mkboot -a "hpux -lq" /dev/rdsk/c1t5d0
# lvlnboot -b /dev/vgroot/c1t5d0 (boot vol)
# lvlnboot -b /dev/vgroot/boot(boot vol)
# lvlnboot -r /dev/vgroot/root (root vol)
# lvlnboot -s /dev/vgroot/swap (swap vol)
# lvlnboot -d /dev/vgroot/dump (dump vol)
# lvlnboot -R /dev/vgroot


Are these commands safe while the production volume group /dev/vgroot is currently in use? The mkboot and lvlnboot commands are performed AFTER performing the vgextend againsts the active volume group.

Do I use the "setboot -p 10/0.3.0 -a 10/0.6.0" command to tell the OS to use the new primary boot disk on autoboot? 10/0.3.0 is the primary boot disk for 11.0. 10/0.6.0 is the path for the older 10.20 primary boot disk. Neither point to the mirrors.

I will come back at a later date to see what I can do to rename the volume group from /dev/vgroot to /dev/vg00. If you can identify these steps also I would appreciate it.

Thanks... jack