RE: RISC Disaster Recovery Disaster

["Pytel, Alexei" <pytela1@xxxxxxxxxxxxxxxx>]

Mike,

What you were describing is the perfect example of what happens when
people neglect to read manuals! I could only wonder how you plan
disaster recovery procedures without consulting Backup & Recovery guide.

When you do scratch install, system will certainly generate new resource
name for every piece of hardware - after all it's a SCRATCH install and
it does not know anything about what was there before.
There's a procedure (described in Backup & Recovery guide) how to
restore entire system image to a system without losing device
connectivity.

Turning off auto configuration was designed for the purpose of not
generating new device descriptions.

The basic sequence is like this:
1)      turn off auto configuration
2)      set console to debug mode - this will guarantee console
connectivity even when device for it is not available (because of 1
above)
3)      system will be up with SRC A9002000 - this is normal
4)      restore everything according to procedure from Backup & Recovery
guide - you better follow correct sequence
5)      resolve resource differences in device descriptions - there's a
procedure for this step as well
6)      set console to normal mode
7)      reIPL - system will pickup all old environment (this IPL is not
necessary - you can just vary on whatever you need, but IPL is usually
simpler)

I've done all this staff dozens of times - it works like a charm. One
only needs to follow documented procedures.


Best regards,
                         
                  Alexei Pytel




        -----Original Message-----
        From:   Mike Shaw [SMTP:mike@mikeshaw.com]
        Sent:   Saturday, May 16, 1998 12:20 AM
        To:     Midrange Listserv
        Cc:     Mike Chwierut; Steve Schmeck; Samir Demir - NAMC
        Subject:        RISC Disaster Recovery Disaster
        Importance:     High

        Hi All,

        Am getting ready to do a disaster recovery test up at our hot
site in Seattle, WA in July.  This will be the first test of a RISC
AS/400 disaster recovery for us.

        To get ready and to see how our recovery instructions needed to
change, I did a scratch install of V3R7 on a Model 40S.  The good news
is that you no longer have a backache from bending over trying to read
the SRC codes to see where it is in the process.  Almost all of the
dialouge is now sent to the system console.  That's nice!

        The real problems began at the point of setting the IPL Options
during the O/S load.  The Set Major Systems Options asks if you want to
Enable automatic configuration which needs to be a 'Y' option.  This
creates a DEVD for the system console and tape devices.  If you say no,
the system hangs with an SRC of A9002000 and a blank console
screen....Some choice.

        By saying yes to enable automatic configuration, I am moving on
to the RESTORE menu to take option 21.  For a tape device in the Specify
Command Defaults screen I enter TAP01 and tell it to prompt for
commands.  This is where it goes to hell in a hand basket.....When
restoring the user profiles I get an error "TAP01 Not Available" and the
restore fails.

        I get back to a command line and do WRKCFGSTS *CTL TAP*.  Auto
config has created DEVD's for all 16 of the 3480 tape drives and the
resource names and device names are not the same!!!!!!!!!!!!!!  We on
the AS/400 side only have 5 of these devices - what has happily been
known as TAP01 through TAP05 for many years (or at least until auto
config was turned on!).  The mainframe has the remainder of the 3480
drives assigned to it.

        Through the process of elimination I was able to determine that
the device TAP01 had a  DEVD created by auto config called TAP06
assigned to the resource TAP01.  I still have to sort through the
remaining 4 drives and figure out which is which....That's not nice!

        Now for the $64,000,000.00:  Is this by design???????????  If it
is, you will never be able to have device naming integrity and have to
manipulate them every time they are created - to include a disaster
recovery scenerio.  <BTW>A disaster recovery site with an auditor
breathing down your neck ain't the place to find out you no longer know
the device name of your alternate IPL tape drive !!!

        I hope someone at IBM will take a serious look at this and offer
an explanation if not a solution to making sure the resource name TAP01
gets assigned to an auto config created DEVD name TAP01.

        My next question is:  Have I missed something here????  Has
anyone else seen this and found a way around this on a RISC box???  Any
comments and ideas would be greatly appreciated.

        Mike Shaw
        North American Mortgage Company
        Santa Rosa, CA






        +---
        | This is the Midrange System Mailing List!
        | To submit a new message, send your mail to
MIDRANGE-L@midrange.com.
        | To unsubscribe from this list send email to
MIDRANGE-L-UNSUB@midrange.com.
        | Questions should be directed to the list owner/operator:
david@midrange.com
        +---
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---