RE: DR Plan - Got one? Will it work?

["Graap, Ken" <keg@xxxxxxxxxxxxxxxxxx>]

Al wrote: 

>Our save strategy is to do a save option 21 periodically,
> and then to do a save of changed objects every day from
> every library, referencing the date/time of the save option 21.
 
>I welcome anyone who would like to challenge this.

Al

What you say could be sufficient, if you had enough time to insure that you
were able to get locks on all your critical objects and that you also had
the time to save them all to tape before someone needed to access them.
People wanting to use the system always seem to get in the way of a backup
process! Now if the system could be quiesced while you did the entire backup
there wouldn't be a problem.

However, in our case we are running as close to 24X7 as possible. I have a
12-15 minute window each night for backing up "changes" and once a month I
can take the system for 1.5 to 2 hours in order to do a full system backup
and IPL. I also have to protect close to 1/2 a terabyte of data.

I don't believe my circumstance is unique either. In a situation like ours,
BU recovery planning becomes much more complex.

During that 12-15 minute window, our system is quiesced in order to remove
all the object locks on our production file set. I backup certain critical
system libraries (job scheduler data etc..) first and then start a Save
While Active save against our production files. When the Save While Active
synchronization has completed I'm able to open the system for use. I am
constantly analyzing the BU procedure in order to find ways I can shrink
this window without compromising the integrity of the backup, because if I
go over my allocated 15 minutes, I will get phone calls from upset users in
the middle of the night!

The backup continues to run for another 3-4 hours and that is only because
I'm able to run concurrent backups using multiple 3590-E11 tape drives.

To insure that I'm continually saving everything necessary for a full
recover, I have also invested in some backup recovery software (BRMS/400 in
my case). I've found software like this to be invaluable when you are at a
recovery site attempting to have the system up and running in the shortest
amount of time. 

We also have a need to recover our Customer Information System to a point in
time as close to the last transaction as possible. To do this we journal all
of our production files and immediately replicate these journaled
transactions to another AS/400 in our network, using the remote journaling
support of OS/400.  

I feel lucky that I have the 12-15 minutes each night. If I didn't have
that, I'd have to resort to some kind of system mirroring scenario. That
isn't necessary quite yet, probably because I've explained to my manager how
much that would cost! Management decided that 12-15 minutes a night was more
"cost effective" ....

The backup is only half the problem though. We test recovering our system at
least once a year. We go to an off site facility and simulate a full
recovery of the entire system from backup. Regular testing has uncovered
several problems that when fixed, made the whole process that much better. I
strongly believe you have to regularly test your recovery process or
everything else is a waste of time and effort. With a good BU plan in place,
recovering the AS/400 in the event of a disaster, should be the easiest part
of any disaster recovery effort. The hard part, especially in this day and
age, is attempting to recover the network which connects all these systems! 

There are many backup recovery scenarios available for the AS/400. With a
little discipline and attention to detail, any one of them may be sufficient
for your particular need. Sometimes I wish I could take the system down each
night and do an option 21 backup, but then I wouldn't get to have all this
fun designing, implementing and testing a complex BU recovery process.
 

Kenneth

****************************************
Kenneth E. Graap
IBM Certified Specialist
AS/400 Professional
Network Administrator
NW Natural (Gas Services)
keg@nwnatural.com
Phone: 503-226-4211 x5537
FAX:    603-849-0591
****************************************



-----Original Message-----
From: Al Barsa, Jr. [mailto:barsa2@ibm.net]
Sent: Monday, April 03, 2000 10:27 AM
To: MIDRANGE-L@midrange.com
Subject: Re: DR Plan - Got one? Will it work?

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@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
+---