1852003 Members
5776 Online
104063 Solutions
New Discussion

Re: make_recovery

 
SOLVED
Go to solution
Ricky B. Nino
Frequent Advisor

make_recovery

Hi,

Tried to create a recovery tape (make_recover -Av) of my system after installation of additional CPU Boards. The creation FAILED.

Potion of the log file repeatedly says...

vgdisplay: Warning: couldn't query all of the physical volumes.
vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c0t6d0":
The specified path does not correspond to physical volume attached to
this volume group

lvdisplay: Warning: couldn't query all of the physical volumes.
lvdisplay: Warning: couldn't query physical volume "/dev/dsk/c0t6d0":
The specified path does not correspond to physical volume attached to
this volume group
lvdisplay: Warning: couldn't query all of the physical volumes.
lssf: Must specify a special_file
usage: lssf special_file ...
save_config: error - unknown disk type for , not SCSI or HPFL
save_config: error - cannot determine root disk

Appreciate any findings and advise...

regards...
Opportunities expand for people willing to put time and effort into learning new skills.
17 REPLIES 17
RAC_1
Honored Contributor
Solution

Re: make_recovery

Is the disk /dev/dsk/c0t6d0 OK?

Has it been assigned in any of the volume group, if yes check pvdisplay.

If it is not then do vgscan and check. You may require to create the lvmtab with vgscan.

Rerun recovery and post.
There is no substitute to HARDWORK
Michael Tully
Honored Contributor

Re: make_recovery

Check to see the status of the disk. It will already be part of an existing volume group. The make_recovery will not run properly or not at all whilst this error exists.

You may need to run a 'vgscan' depending on the outcome of the disk problem. Let us know what you find.
# mv /etc/mnttab /etc/mnttab.orig
# vgscan -a
Anyone for a Mutiny ?
Norman_21
Honored Contributor

Re: make_recovery

Hi,

Add to your notes, to upgrade your ignite utility. download it from here:
http://www.software.hp.com/products/IUX/download.html

This is more effecient than make_recovery. The new command will be :

make_tape_recovery -x inc_entire=vg00 -I -v -a /dev/rmt/0mn

Thanks & Good luck.
"Attitudes are contagious, is yours worth catching"/ My first point was given by SEP on January 31, 2003
T G Manikandan
Honored Contributor

Re: make_recovery

Just check the /etc/lvmtab for the disks for the VG.

ALso do a vgdisplay on that VG and check the number of disks and the information of c0t6d0.

If the c0t6d0 is not available in the vgdisplay list of the VG then you can do a

#mv /etc/lvmtab /etc/lvmtab.old
#vgscan -a

Revert
Steven E. Protter
Exalted Contributor

Re: make_recovery

1) make_recovery is obsolete.

Start over with make_tape_recovery

This should be done if you want support on disaster recovery in the long term.

2)save_config errors can be caused by the new Veritas Volume manager. If you are not using the new volume manager you can safely remove the volume manager and eliminate some of your errors. That's what the support center had me do when I ran into these problems a few months bacik.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Ricky B. Nino
Frequent Advisor

Re: make_recovery

Guys,

Just to give you further details...
vg00 LVNames...

I checked on the following

1. /etc/lvmtab. /dev/dsk/c0t6d0 is included under vg00.
2. /etc/mnttab. /dev/vg00/lvol2 is missing
3. vgdisplay -v vg00. Information related to /dev/dsk/c0t6d0 is missing.

Running #vgscan -a will recreate /etc/lvmtab alone. Please correct me NOT so sure about this...

How will #vgscan -a affect the content of /etc/mnttab and the output of vgdisplay?

thanks in advance...
Opportunities expand for people willing to put time and effort into learning new skills.
T G Manikandan
Honored Contributor

Re: make_recovery

you are right.

If vgdisplay is missing that disk in the VG then it would be like the disk was removed from Vg but not properly updated in the /etc/lvmtab file

yes just do a

#vgscan -a

that should re-build the lvmtab file.

Then lvol2 in the VG00 is the swap volume which will not be displayed in the mnttab.you can use swapinfo -ta to check that it belongs to swap.

Thanks


John Dvorchak
Honored Contributor

Re: make_recovery

You have a corrupt /etc/lvmtab file. Move /etc/lvmtab to save it and then run vgscan -a to create a new one. I have had this exact same problem before and someone on the ITRC gave me the fix for it.
If it has wheels or a skirt, you can't afford it.
Ricky B. Nino
Frequent Advisor

Re: make_recovery

Guys,

Tried to execute #ioscan -fnC disk. The problem disk c0t6d0 was NOT listed.

FYI...
Opportunities expand for people willing to put time and effort into learning new skills.
Norman_21
Honored Contributor

Re: make_recovery

Did you check to see if the disk is connected and if the SCSI settings is configured properly. Check the dev lssf /dev/*
Then do insf -e.

Hope this help.
"Attitudes are contagious, is yours worth catching"/ My first point was given by SEP on January 31, 2003
Ricky B. Nino
Frequent Advisor

Re: make_recovery

Guys,

Is there any other way to get info on this disk aside from "ioscan"?

Looks like I have a h/w failure in here. Though I was STILL able to access what's on the disk.

I am thinking of rebooting the system. But really worried that it might NOT get the system up anymore...

regards...
Opportunities expand for people willing to put time and effort into learning new skills.
T G Manikandan
Honored Contributor

Re: make_recovery

I do not see any problem for panic.
Anyhow post the output of /etc/lvmtab and the output for vgdisplay -v /dev/vg00.


REvert fast
Ricky B. Nino
Frequent Advisor

Re: make_recovery

Hi Revert,

see attached. also included the output of "ioscan -fnC disk".

btw, many thanks for looking into this positively...

regards...
Opportunities expand for people willing to put time and effort into learning new skills.
Michael Tully
Honored Contributor

Re: make_recovery

Hi,

The disk belongs to /dev/vg00 and as you can see from the vgdisplay output the number of disks list in
CUR PV=6
ACT PV=5

is where the problem lies.
okay see in your output, where it states, that you have 'available/stale' for some volumes, well this is the mirror of part of your root volume group.

Get the disk replaced. Seeing it is mirrored, the system should fix the mirror itself depending on the model of your system. Some are hot pluggable and some are not.

Good luck
Michael
Anyone for a Mutiny ?
T G Manikandan
Honored Contributor

Re: make_recovery

It seems like there is quite work to be done.
Make sure that you backup things before proceeding on the same.

The number of PV's in the VG
Cur PV =6
Act PV =5

The problem is this.

If this would have been for a non VG00 then a
vgreduce -f would solve the problem.

As this is for vg00
you need to

Boot into LVM maintenance mode.
Interrupt the boot process then
hpux -lm

Export the vg00
vgexport -m vg00

Re-create the VG00
mkdir /dev/vg00
mknod /dev/vg00/group c 64 0x000000

Again do a vgimport for vg00
vgimport -m /dev/vg00 pv_path
In this process if there are any errors then just do a
vgreduce -f /dev/vg00
mv /etc/lvmtab /etc/lvmtab.old
vgscan -v

Now the vgdisplay -v /dev/vg00 should have Cur PV and Act PV with equal no. of disks

Reboot
/sbin/reboot


you might face problems during a make_tape or during the addition of disk to this VG.

Just plan a downtime for this accordingly.

Thanks
T G Manikandan
Honored Contributor

Re: make_recovery

I would also like to have the following information like

#ll /etc/lvmconf/vg00.conf
#vgcfgrestore -n /dev/vg00 -l

Also Is your vg00 mirrored?

Revert on the same
Ricky B. Nino
Frequent Advisor

Re: make_recovery

Guys,

Thanks to all.

Ill have HP Engineer here to replace the defective disk.

Thanks for all the inputs. Appreciate every input you've shared.

regards...
Opportunities expand for people willing to put time and effort into learning new skills.