make_tape_recover vs. make_net_recover

[John Jimenez]

Hmmm.... because I do not have a server, does that mean that I have to use make_tape_recover to my LTO1 Library over the SAN? When using the make_network_recovery is the ignite server usually a dedicated server? What are the requirements? I currently one RP7410 running 11.11 (which normally uses dds2 and takes quite a long time) and the 2 new ones running 11.23.

[John Jimenez]

a few years ago a consultant said that we could not use the LTO1 tape library/drive over the SAN to make_tape_recover, only dds tape drive. But I just ran it to the first LTO drive and if ran fine and of course way faster then that old dds drive.

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

Like I had mentioned I have never had to restore from a make_tape_recovery. is doing it this way accepted? Will I be able to restore from this if I need to?

Thanks again for all your help

[Torsten. · Accepted solution]

Hi John,

regarding the restore over SAN:

The firmware of the server is not clever enough to find the tape device inside the SAN and boot from it.

For this reason you need to boot from the install media (DVD) and select the tape device inside the SAN, because the install utility from the media *is* clever enough:

-> Install HP-UX -> select Advanced Installation -> Recovery tape -> GO!

You can also install a current version of ignite on one of your servers and set it up (see the link I gave you earlier). Ignite is able to handle both, 11.11 and 11.23.

http://www.docs.hp.com/en/IUX/download.html

The manual (Ignite-UX Administration Guide: for HP-UX 11.00 and 11i - Edition 23, June 2006):

http://docs.hp.com/en/B2355-90970/index.html

[Bill Hassell]

When is a tape drive not a tape drive? When it is in a SAN... A SAN is nothing like a simple SCSI cable to the tape drive and the processor ROMs are far too small to contain all the specialized code to talk to a storage network. The LTO drive (as you have seen) is just fine for Ignite. What the consultant may have meant was that older computers may not have firmware that knows about LTO drives. Yes, there are a lot of differences between DDS, DLT, LTO, AIT and other tape drives. But your rp's will work OK with SCSI tape drive (for recovery). You may need to get a portable LTO drive just for recovery (and process testing) tasks.

[Michael Steele_2]

Yes you'll have to use make_net_recovery but there are cons to this. For example your ignite server will have to be rebuilt in the event of a disaster before your production servers. Some NICs don't like igniting over a subnet so you'll have to verify this. You can go through a router but you'll need boot helper to do so and that's and additional complication. In short, your adding extra steps to your disaster recovery plans. So its my preference to use tapes unless you're in a server farm. Tapes are uncomplicated and easier to manage. Consider adding a dat drive into your servers and simplifying your duties.

Maybe I'm old fashioned but I personally like have a tape on my desk and one in the vault.

[John Jimenez]

Thank you for all that information. I used the make_tape_recovery not the make_net_recovery and it finished okay. Since I have never restored off one, I will use the one with less steps, which is tape. I now have a D.R. system off site (K380, connected via a SCSI not SAN,to a MSL5030 Library, I will have to plan some time to play and see if I can. I will be assigning points in a little while. Are the links that are posted the best links to learn how to do the restore? I will try and see if I can make time this week and next to do the restore on the DR system off site.

[Torsten.]

The link is the latest official manual as mentioned:

The manual (Ignite-UX Administration Guide: for HP-UX 11.00 and 11i - Edition 23, June 2006):

http://docs.hp.com/en/B2355-90970/index.html

[John Jimenez]

Thanks again everyone

[Michael Steele_2]

If you've never used a make_tape_recovery tape then you should find a test machine and do so. You can choose an automatic restore which takes about an hour and one half or interupt the auto restore and make O/S config changes. This is the clone process. And for an off site disaster restore you'll have to use the non-auto restore since some things will be different. The gateway ip for example.

Also you can verify the make_tape_recovery tape with tar and you should include this after every creatation.

mt -t /dev/rmt/0m rew (* verify the tape is rewound
mt -t /dev/rmt/0mn fsf 1 (* move the tape ahead one record and past the boot lif
tar -tvf /dev/rmt/0mn (* list the table of contents

and selective restore of only one file

cd /dir
tar -xvf /dev/rmt/0mn dir/file_to_restore

[John Jimenez]

Micheal,

Thanks for pointing me in the right direction. This really helps!!
The D.R. system at the Warehouse is my test system.