Re: Switching ID's -- MIDRANGE-L

Al,

Regarding the risk analysis, I agree. I've always worked DR plans basedupon scenarios and categorize each one; from as simple as Workstation_Agets toasted to a hurricane just blew in the front door and left out theback door. From as semi-complex as my boss just kicked out the powerplug leading to the UPS (which he did) or a backhoe operator just cutall of our phone lines.

The scenario that I was tackling with this question was: The buildingand its contents are fine; there's power; the production server just bitthe proverbial dust. As reliable as the iSeries is, I once worked at aplace with six of them, but there was one that insisted upon going southon us (disk drive, no RAID) periodically.

And a test is definitely part of the plan. We live off of our phonelines via which orders are transmitted directly to our iSeries sore-directing the comm cables to SystemB is a must. And the hurricanescenario becomes a real pain.

Threshold of pain is the criteria/question that I pose to managementwhen they tell me that an alternative is too expensive. Here,apparently, an HA package is more painful than shutting down thebusiness. But I understand what you're saying about currency ofbackups. By 0700 our backups are out-of-date so its either journaleverything remotely or push mirror images of the transactions to backupfor a post-restore update. I once tested users' adherence to the paperflow requirements by telling them that the system was down and that adisk drive had to be replaced; they were supposed to pull theirpaperwork for re-entry. Not one person was able to comply (which Ireally knew would be the case) even though there were procedures inplace for this.

As far as the multiple iSeries icons, that is not a problem as regardsthe users; connections have only been established to production(SystemA). Only IT (my boss and I) have access to both systems on anormal basis. However, if I have to divert users to the backup system,I want it to be seamless. I.e., once the switch has been made, tablesupdated, etc., all I want them to have to do is double-click that sameol' icon on their desktop.


Thanks for all of the suggestions.


        * Jerry C. Adams
*IBM System i5/iSeries Programmer/Analyst
B&W Wholesale Distributors, Inc.* *
voice
        615.995.7024
fax
        615.995.1201
email
        jerry@xxxxxxxxxxxxxxx <mailto:jerry@xxxxxxxxxxxxxxx>



Al Mac wrote:

I would suggest that there be some RISK ANALYSIS ... I think the odds ofsomething going seriously wrong with the 400 are small compared to the riskof the site being hit by disaster. Depending on your region of thecountry, and other factors, the risk will vary for what can come from badweather, earthquake, medical epidemic, criminal intruders, terrorism,insider misbehavior, vehicle crash into your building, etc.


When you think you have all bases covered, you need to schedule a TEST.
Either periodically thereafter, or after various upgrades, again TEST.

You may have package software that is licensed to SystemA CPUserial#. They may be willing to provide software key for SystemB at low cost.

Between time of last backup and the crisis, there may have been input workthat needs to be redone. Something to think about is paperwork flow suchthat you can reconstruct what was done since last backup.

Some package software keeps track of user input not by user-id but byworkstation-id


A PC can have more than one desktop icon for 400 access.
One for SystemA, another for SystemB,
which might want to use same workstation-id logic on both systems.

We have two [2] 520's.  One is production; the other is our backup
system.  Both are single partition.  There are 20 - 30 workstations and
about the same number of printers.  The workstations are primarily PC's
using iSeries Access for Windows; we have a few twinax workstations
(including the system console).


So far we (knock on wood) haven't had to use the backup system for
production.  Someone had the radical idea, however, that we actually
needed a disaster recovery plan.  I need a little (?) help with an
aspect of the plan.


For this phase we are assuming that the damage is isolated to the
production box (SystemA); that there is no structural damage to the
environs; that everything else is copasetic.  I port the configuration

from SystemA to SystemB (the backup) in a save file (SAVCFG)

automatically every night.  (Forget to mention: Both boxes are supposed
to be identical in all regards.)  Security data (SAVSECDTA) is, also,
pushed to the backup in a save file.


My question relates to user access via iSeries Access for Windows.  Each
PC user's IAW is configured to identify SystemA as the system to which
they will access programs and data.  Even though there are "only" about
20 PC's attached this way, I would rather not set up a connection to
SystemB and, then, change each PC to access the backup system.  Instead,
I would rather change the identifier on the system (520) itself.


It looks like CHGNETA can be used to rename the systems.  Is this all
that I need to do?  Or do I, also, need to change the IP address for the
box?  I seem to remember that there is a server somewhere that says, "IP
address 10.0.0.4 = SystemB" etc.  Is that the Domain Name Server (DNS)?
And would I need to change that, too?


Thanks.
--
       * Jerry C. Adams
*IBM System i5/iSeries Programmer/Analyst
B&W Wholesale Distributors, Inc.* *
voice
       615.995.7024
fax
       615.995.1201
email
       jerry@xxxxxxxxxxxxxxx <mailto:jerry@xxxxxxxxxxxxxxx>