>We've been doing DR mirroring for 20 years. It gets tested often. We've moved >production twice to another >data center using our procedures. What we've >never done is run production in another location >temporarily. 'Temporary' >means move it, run it until at least one transaction is committed, then move >it >*all* back. That is hugely complex and costly.
>A lot of management fantasizes about a big A-B switch that we throw one way or >the other. So wrong. About 5-6 years ago, I was working as a vendor (Daily Support) for Credit Card Software for Halifax/Bank of Scotland. I believe the first time I was given a "heads up" was when we were supporting 12 Million Cardholders. The "Heads up" was that on Friday evening at 6pm Main site was going to shut down and the whole weekend PRODUCTION was going to run on DR site, then 6am Monday Morning, Main site is to come back up and continue as if nothing happened !!! I literally peed in my pants !!! Probably everyone in Atlanta could hear me...NOOOOOOOOO !!! Thinking of all the network signons with Visa and Mastercard...all the credit card Authorizations...there was absolutely zero chance of this working without issues. Well, came in Monday morning after receiving no calls over the week-end ....everything was fine. We ran the Batch Monday night...and that would be pulling in transactions from over the week-end from DR site. NO ISSUES !!! Everything was fine. Whoever did this, from a Systems perspective....I tip my hat. I've never seen someone do this with Production, but it worked fine...so what do I know. Never seen anyone else do this in my 42 years of Mainframing either. Unfortunately, Halifax/Bank of Scotland is no longer with us. They were absorbed by Lloyds Banking Group. Thanks, Tom Savor ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
