Disaster recovery: Copy savesets from tape to disk first?

[goony]

My current Networker server hardware: Sun box with dual tape drives + 750GB of disk space (Legato adv_file) used for backups. I make backups (full + incr) to disk, then 'clone' the savesets to tape to be taken offsite.

Next month I will be doing a 'disaster drill' in which a group of offsite tapes will be taken to another system for restoration. At the alternate site I have a Sun box with 750GB disk storage, very similar to the original server.

What I would like to do is restore my original networker server via the usual recovery steps, but then clone all of the savesets from each tape to the disk storage - this will permit me to be able to do multiple client concurrent restores and reduce contention for tape drives and volumes.

One twist: The offsite tapes have about 25 client systems on them, but I only want to restore 10 selected clients at the disaster site.

My current thinking is to do the following steps:

1) Create a 'clone pool' and mark my adv_file disk device as being part of that pool

2) For each tape volume that I have at the disaster site, build a list of savesets:

[tt]mminfo -c <client name> -r ssid >> savesets_<volume_name>.list[/tt]

Do above step for each client I want to recover until I have a list of all savesets on tape <volume_name> for all clients.

Repeat above for next tape volume... eventually, will have a saveset list for each volume; each list will have the ssid's for all 10 clients.

3) Put tape <volume_name> into drive and mount it under Legato networker

4) Use this command to clone desired savesets from tape to adv_file disk:

[tt]nsrclone -b 'clone pool' -S -f savesets_<volume_name>.list <volume_name>[/tt]

Repeat step 3 & 4 for each tape volume.

Will the above steps accoplish what I want to do? Once cloned, Networker should happily use these savesets for restores since they are all 'online', right?

Thanks!

Goony

 

[605]

In general, your solution will work. But keep in mind:

1. The system you will do a DR of the NetWorker server must have the same hostname.

2. By default, NW will try to recover from the original save set (actually the one which has been written first). To force NetWorker to recover from a clone you can either set the original save set to "suspect" or use the recover command which allows you to specify the cloned save set via the cloneid: recover -S ssid/cloneid [-a pathname]|[files]recovNetWorkerTo build a "save set list of a volume" use
mminfo -q "volume=vol_name" -r "ssid"

3. Do not select devices for a special pool if not necessary, in general.
This can cause a lot of problems.
But on a file device, you can avoid this easily: Just make sure that no media for this pool can be found anywhere else.

4. Especially, do NOT select a device for a pool when using file device types!
While tapes are "movable" (can be receovered from any device at the server/restore node), file devices are not.
You must use the exact device name as you used before. Unfortunately, this is not always possible and the
user often does not remember.

Yes, this info is of course in the resource files. However, if you can not restore these from an (A)FTD due to a missing device name, you will be in trouble.

So i would always prefer to restore my bootstraps (and index) from tape and not from (A)FTD.