Disaster Recovery Build Scripts

[Allan Bateham]

I test disaster recovery in an off-site, third-party facility. I use ignite to restore the base system, however, since the systems are dissimilar, it requires that I rebuild my non-vg00 volume groups, logical volumes and file systems.

I have a 8 page document that is a shell script the supposedly automates this process.

Since I am not a Unix pro, I do not understand the script 100%. And since the document is 8 pages long, retyping it is error prone.

I was wondering if someone could point me too a script that is downloadable and, preferably, well documented.

[The Real MD]

First things First

get yourself on a sys admin course for hpux, no amount of scripting can account for everything for every situation. But the knowledge to understand and overcome problems will be far more useful.

http://hpux.cs.utah.edu/hppd/hpux/Sysadmin/sysinfo-3.3.1/

use the program from the above link to document your system. From which you can recover your system.

I hope this helps

Martin.

[Allan Bateham]

Thank you Martin, I have sysinfo and maintain current hardcopies.

What I am trying to accomplish is streamlining the recovery process and avoid as much manual steps as possible.

[BRUEL Pierre]

Allan,

I'm not sure that an automated process is the best solution for DRP.
The only thing which will change on the DRP site is the creation of VGs, since the disks used to rebuild them won't have the same sizes than yours. Building these VGs manually don't take much time.
For LVs and FS, you can prepare a script on your site with all commands for the different LVs and FS :
lvcreate -L -n /dev/
lvcreate -L -n /dev/
....
then :
newfs <-o largefiles> /dev//r
....

Regards,

[Mark Grant]

Allan, a generic script for this is going to be very difficult to find. However, there are some steps you can take to reduce the DR downtime.

Do Ignite backups to DLT drives or SDLT if it's supported. Have a primary machine that is set up as a DNS slave, a NIS master/slave (if you use it) and if you can, have your entire backup software catalogue contained within vg00. Get as much infrastructure services in vg00 on one machine. It doesn't have to be the only source of these services, for example DNS could be on Windows, NIS master could be another machine just as long as this one machine is set up to serve these services when the primary servers are not available.

IF you do this, a large part of the DR headache is done for you by the initial Ignite restore. If you can squeeze an Ignite server in their too, you could have a simple script that recognizes you are in a DR situation and builds all your other DR machines and then does the entire restore for you from here staright from the initial Ignite restore.