RE: Two-way mirroring -- MIDRANGE-L

I agree and have had this lingering over me for years.  We have three as400
performing load balancing and D/R on our main stay of business.  Transaction
come from all over the nation.  The reason it works for us is the fact than
95% of the activity are new transactions.  The other 5% are updates but the
application copies the original transaction marking it as deleted and
creates a new transaction.  This way we see the before and after.  Our
biggest problem has been having 3 users using three different applications
to update the same record from the same file at the same time, we do not
lock the records while the programs are waiting for user input.  This is why
we created an update module that received the before and after image from
application, reads the record with the lock, compares the before to the
after and if changed, compares the before to the file.  If there is no
difference, the after gets moved to the file.  If there is a difference, we
check the after to the file, if the same no action is needed, if different,
an error code is returned to the application.  Works great, we get very few
errors returned.

Christopher K. Bipes     mailto:ChrisB@Cross-Check.com
Sr. Programmer/Analyst   mailto:Chris_Bipes@Yahoo.com
CrossCheck, Inc.         http://www.cross-check.com
6119 State Farm Drive    Phone: 707 586-0551 x 1102
Rohnert Park CA  94928 Fax: 707 586-1884

*Note to Recruiters
Neither I, nor anyone that I know of, is interested in any new and/or
exciting positions. Please do not contact me.


-----Original Message-----
From: Terry Herrin [mailto:therrin@isaac.net]
Sent: Thursday, November 04, 1999 5:57 PM
To: MIDRANGE-L@midrange.com
Subject: Two-way mirroring


"Edward Koziol" <koziol@chaney.net> wrote:
>They can do it and so can DataMirror...

It would be interesting to know how these products handle multiple
updates to the same data.  For example, if system A updates a field in
a record with value "X", and system B updates that same field and
record with value "Y", what value are you going to use when you go to
resynch the databases later?  Are you going to compare the journal
entries from both systems to determine which update happened later,
and then use that value?  Or what if system A deletes a record that
was updated by system B?  Do we just delete that record and not bother
with why system B was doing an update to it?  Whatever choices are
made in the above examples, assumptions will be made according to a
set of rules.  How am I to know those rules are the way I want my data
to behave?  This sounds like it would be a great thing to have, but to
me it opens up a whole Pandora's box of issues.

Terry Herrin
Sr. Programmer/Analyst
New Hanover Regional Medical Center
Wilmington, NC
+---
| 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
+---