Ignite-UX make_net_recoverey restore problem

[DD-CC]

Hi There

I'm testing my ignite setup before implementing it onto out live systems, and I'm having problems restoring to a clean set of disks - if I'm restoring onto the running system everything works fine.

The Ignite client is configured with 1 root disk (vg00) and 3 datadisks (vg01 and vg02). It's an older HP9000/715 running hpux 10.20.

I'm using the comand: 'make_net_recoveer -s -v -n 6 -x inc_entire=vg00' to make the backup.

I want to test a restore onto a system with new disks and to do this I remove everything (lvol, vg, pvg) from the datadisks (vg01 and vg02) so that they are displayed as 'unused' by SAM.

When I recover I get the following errors :

* Executing user specified script:
"/opt/ignite/data/scripts/os_arch_post_c".
* Running in recovery mode (os_arch_post_c).
* Executing user specified commands.

vgimport: Warning: Volume Group contains "0" PVs, "1" specified. Continuing.
vgimport: Quorum not present, or some physical volume(s) are missing.
Logical Volume is not defined on any physical volume.
"/dev/vg01/lvol1" is missing Physical Volumes.
Beginning the import process on Volume Group "/dev/vg01".
Volume group "/dev/vg01" has been successfully created.
Warning: A backup of this volume group may not exist on this machine.
Please remember to take a backup using the vgcfgbackup command after activating the volume group.
vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c0t4d0":
Cross-device link
vgchange: Warning: couldn't query physical volume "/dev/dsk/c0t4d0":
The specified path does not correspond to physical volume attached to
this volume group
vgchange: Warning: couldn't query all of the physical volumes.
vgchange: Couldn't activate volume group "vg01":
Quorum not present, or some physical volume(s) are missing.

ERROR: Executing the commands "
/usr/sbin/vgimport -v -m /etc/lvmconf/vg01.mapfile /dev/vg01
/dev/dsk/c0t4d0
/usr/sbin/vgchange -a r vg01

... and the restore fails.

The root disk is ok, but the datadisks are only halfway configured, and I haven't succeded in restoring vg01 and vg02 without stating over again.

Can anyone help me here ??

Thanks in advance
S?ren Dreyer

[Stefan Farrelly]

Hmm, tricky one. I think the key errors are;

vgimport: Warning: Volume Group contains "0" PVs, "1" specified. Continuing.
and
ERROR: Executing the commands
/usr/sbin/vgimport -v -m /etc/lvmconf/vg01.mapfile /dev/vg01 /dev/dsk/c0t4d0

This implies that the vgimport of vg01 from c0t4d0 doesnt match the mapfile vg01.mapfile. Do a strings on the vg01.mapfile, how many disks are specified in there, should be one, whereas the first error above indicates there are no volumes specified in the mapfile.

[DD-CC]

Hi again,

and thanks for at quick answer.

There is no disk defined in vg01.mapfile, only a logical volume !
# strings /etc/lvmconf/vg01.mapfile
1 lvol1

... vg00.mapfile dosen't have a disk defined there either.

but in /etc/lvmtab the disk is defined :
# strings /etc/lvmtab
/dev/vg01
/dev/dsk/c0t4d0
#

S?ren

[Stefan Farrelly]

Ignore my last reply!

the map file is supposed to only have the lvols defined in it, on the vgimport command line you specify the vg name and disk its on. Hence, when you restore back to the system you made the recovery on, it will work fine. The disk is there with the same name (c0t4d0) and it can import it fine.

However, if you use this tape to restore to another server then when it goes to import vg01 from a disk called c0t4d0 (if such a disk exists with the same name) but of course it cannot import it becuase your vg01 data is not on that disk (its back on the disk of the same name where you made the recovery from!).

So the approach youre using isnt correct. The normal procedure is to only have the make_recovery image contain vg00, and if you have other vg's then they are restored via a normal backup procedure after youve restored vg00 and created the other vg's you need. Make_recovery is designed to make a bootable backup of VG00 only (or if say /usr or /opt is on another vg, eg. vg01, it will then include vg01 also).

[DD-CC]

Hhmmm, see your point

I guess my understanding of Ignite wasen't correct.

I know that Ignite is for making backup of the root volume group (vg00) and not databackup. But all your vg definitions and configuration of the other volume groups (vg01, vg02 etc.) is placed as configuration files on the root vg.
So when a fire destroys my server (or alle disks crashes somehow), and I can't remember how my disks was configured - how can I recreate the disk configuration on the server?
I thought that Ignite could do that. By replacing all the disks and do a ignite recover, the root vg (fully restored) and disk configuration would then be ok - and ready to restore the normal backup onto.

Is the only solution here to have the disk configuration written down somewhere ??

[Stefan Farrelly]

Ignite will restore the files for the other VG's, but not the vgs themselves. You can then manually re-import your vgs if you know what disks they are on or use vgscan to do it, but even better use ignites print_manifest command as it gives a great list of your disk/vg layout for use in disaster situations. We run this command regularly and stick it on an external web server so its nice and safe.

[John Palmer]

It's always good practice to have your LVM configuration documented. There are several scripts available to help you - do a forums search and you'll find them.

I always script the creation of volume groups and volumes and keep them in a directory (/disaster) that is easy to recover from tape.

Disaster recovery is then a matter of recovering your root disk, vgexporting your data VG's to get rid of them, tailoring the scripts in /disaster to cater for differences in hardware layout (disk sizes and paths), creating the data VG's and volumes and finally recovering the data.