Folks,

I apologize if this is a FAQ....

(I cannot find a "complete" documented procedure)

:0(


Currently, we use the following procedure to restore systems:

1) Restore the OS using the "cloaner" method as published by IBM:
   a) Boot off of the AIX installation media
   b) Restore system from tape (mksysb)
  (This allows you to recover the OS to dissimilar hardware)
2) Reconstruct the network (including the TSM backup network).
3) Restore all application data (and system data if needed) via TSM.

This methodology does not support a true DR test as our mksysb tapes are
not cut regularly or vaulted.

Unfortunately we do not have a solution that allows us to cut bootable AIX
tapes on a regular basis.  (Not all of the systems have local tape drives.
The procedure can not be automated. etc.)

In the future we must restore all RS/6000s with only our vaulted
installation media (OS and TSM) as well as our vaulted TSM tapes.  Due to
this we must develop procedures for a total bare metal restore of an
RS/6000 using only the OS installation media, TSM backup tapes, and any
relevant TSM disaster recovery reports.

We would like to develop a procedure that details the recovery of a generic
RS/6000 using only installation media, TSM backup, and disaster recovery
reports.

Ideas/Things to consider:

We can not assume that we will recover systems to identical hardware.  As
such we can not simply install a base OS and restore.
 What system files (/etc/passwd, /etc/hosts, etc...) are relevant to the
recovery of the system.
 What system files should not be recovered?  (I.e. kernel, driver specific,
etc.)


Can anyone help with any existing procedures or point me in the proper
direction?

Thanks!


DaveZ