Disaster Recovery preparedness

[Joe Profaizer]

I'm looking for a good checklist of disaster recovery preparation for our HP-UX L2000 servers. For instance,
Filesystem listings
Ignite tapes
etcc.............

[Santosh Nair_1]

In addition to ignite images (tapes or server based), we also maintain a list of the various servers' specifics, i.e. model, number and type of cpu, memory, number of disks attached, LVM configuration/filesystem layouts, kernel parameters, ioscan listing, cronjobs and defined printers.

I have a bunch of scripts that I use to gather all this informatino together...then I found out about the print_manifest command that comes as part of Ignite that does essentially the same thing.

-Santosh

[Edward Alfert_2]

HP has an entire document tree on disaster recovery:

http://us-support2.external.hp.com/iv/bin/doc.pl/sid=4d88bd39140f6d93df/screen=ivHome/?NODEID=English_SW::WW_SW_UX_HAT_EN_E/Q1.3.1&WARP=1

[Bernie Vande Griend]

Physically: ignite tapes, backup tapes, code words and license keys (you could get these later, but it will save time if you have them handy), and disaster recovery plan (what do you do in a disaster, what is most important, who do you contact, etc.)

Information:
Ignite's print_manifest is a great start to gathering information you'll want.
You can also use scripts cfg2html, nickel, and/or sysinfo to gather even more information. These 3 also write the information to html. These html files can then be saved to tape that can be stored off-site. There are links to these scripts in the forums. You should also have serial numbers, product numbers, and part numbers for each piece of equipment. The scripts won't be able to get these unless you put them in a file on the server (Ignite has a way to store it too).

In a disaster, you'd want as much information about your system as you can get, so it doesn't hurt to have several sources of it.

[Michael Tully]

Hi,

The list is just about endless in a disaster
recovery situation depending on what areas
you wish to focus on. For a HP-UX server these
are what I would have. And some questions that
your management have to decide on like

What constitutes a disaster?

For every server that I have, I collect the following
information and record in both hard and soft formats,
with a copy remianing off site.

MAIN SYSTEM INFORMATION
VOLUME GROUP OUTPUT
Volume Group vg00
/etc/fstab
OTHER HARDWARE
ioscan output
SOFTWARE swlists, applications Whereabouts of CDs

OPERATING SYSTEM CONFIGURATION & PARAMETERS
Kernel (System file and boot.config)
lsdev
sysdef
/etc/services
/etc/inetd.conf
/etc/inittab
Run Level start/stop
/etc/syslog.conf
Swap & Dump Information
Lan related parameters
Printer Information
User Administration
Crontabs
Have scripts that you can easily re-build your LVM
structure. This is necessary if the disks are not
compatible with the ones you use now.

Build an initial HP-UX operating system using the production server backup created from your make recovery tape.
? Check that the system will boot from the kernel created from the existing production server.
? Initialise discs and build the required LVM structure. Have a copy of vgexport files for each VG.
? Restore application software, system tools, local utilities and client data.


My 2 cents worth.
Michael

[Jeremy_7]

This is a helpful page post by hp.

http://docs.hp.com/hpux/pdf/B2355-90704.pdf

[Scott Van Kalken]

I'll throw in my two cents worth as well here.....

In addition (again) to everything else that people have - which is all good. I like to have a cstm listing of hardware.

# cstm
cstm> select device all
cstm> map
cstm> info
cstm> infolog


write to a file etc and print out.

I like to have this just in case of fire or whatever.

I also have like a "fact sheet" for each machine that has most of this stuff as well as contact numbers for all of the key people involved with each system (i.e. DBA's Application support, my boss etc...)