DR from make_tape_recovery

[Chushia]

I performed a DR test with ignite make_tape_recovery tape. The tape was made on vg00 of source machine. The source has vg00 (2 disks) and vg01 (20 disks).

On the target machine, I restored vg00, configured vg01.

Then I reboot the target server, now vg01 did not recognize the disks anymore.

I noticed ioscan gives different device names. For example, it was /dev/dsk/c12t0d0, after reboot, it changed to /dev/dsk/c4t0d0 ......

Why is this happening? What is the best practice to perform a system recovery on a different but similar hardware?

[Dennis Handly]

>system recovery on a different but similar hardware?

What were the two types of machines? What OS version?
Do you have the same number and types of disks?

[Chushia]

The source machine is rp5430, target is rp8400. OS is hp-ux 11i. Both machines have the same number of disks (2 internal, 20 external) and are all 36G each.

[James R. Ferguson · Accepted solution]

Hi:

> The source has vg00 (2 disks) and vg01 (20 disks).

Ignite is intended for the recovery of vg00. By this, I include the recovery of vg00 in a disaster recovery on alternate hardware. One recovers their operating system (vg00); configures non-vg00 volume groups from available disk; and restores from backup the non-vg00 data.

> On the target machine, I restored vg00, configured vg01.

I'm not sure what you mean here. The Ignite Reference Guide states "If any file from a disk or volume group is included in the recovery archive, that disk (or all disks in the volume group) will be reformatted during the recovery, and only the files included will be recovered. Any files that were not included in the archive, will have to be restored from normal backups. Disks and volume groups that did not have any files included in the archive are not reformatted during a recovery and are reimported and remounted at the end of the recovery."

> I noticed ioscan gives different device names. For example, it was /dev/dsk/c12t0d0, after reboot, it changed to /dev/dsk/c4t0d0 ......

This would be expected particularly if you are recovering to different hardware. This can also happen if you clone to the same hardware.

> What is the best practice to perform a system recovery on a different but similar hardware?

See my opening comments.

A better description of what you mean by "a DR test" would be nice. Did you Ignite to a different server or overwrite the one from which the Ignite tape was made? Did you choose the Advanced Installation mode and toggle "Import Volume Groups?".

Regards!

...JRF...

[sujit kumar singh]

hi,
make_tape_recovery is very hardware dependant so you can not be sure if there is a defference in the hardware paltforms of the source and the target servers.
better that you can take the self make_tape_reovery and restore the same on the self server that means having the independent make_tape_recovery media for the individual servers.

Regards
Sujit

[Chushia]

Thanks for your replies.

On the source server I did "make_tape_recovery -x inc_entire-vg00 -I -v", and on the target server I did Advanced installation and had set IP and subnet.

After the target server was up and running, I exported vg01. Based on ioscan result, I did pvcreate for disks available, then created vg01 with these disks, and restored other data onto vg01.

Then I did a reboot, and could not access vg01 anymore because the devices were changed.

Then I got stuck. I don't quite remember now, but it seems I could not create vg01 nor remove it ...... quite a mess.