Patrick,
With web servers no - you will face no downtime as it is only a sort of a 'portal' if you will with no user developed or customized code. The downtime with a web server could well be as minimal as a user connected to an old instance who might get kicked out so may have to log in again. I think this discussion started with code migration. Downtime there could happen during an upgrade or migration where you can safely assume that roughly about 20% or above of the code will change, data will need to be migrated, user caches would need to be rebuilt, a possible DNS refresh might need to happen in case of server name changes, etc. Joe PS: Apologies to those who might have responded to my responses to you'll and have not heard from me - I am reading all these mail in a reverse chronology to catch up with some mails I missed. _____ From: Action Request System discussion list(ARSList) [mailto:[email protected]] On Behalf Of patrick zandi Sent: Wednesday, March 19, 2014 10:21 AM To: [email protected] Subject: Re: BMC Remedy and Flash ** Joe help me understand the downtime with a server group scenario. If you have 3 web servers (which can rotate each out and be patched, os and application wise - then rotated back in), and have 3 Servers is a server group (which again be pulled patched, upgraded, OS patched and put back into group) one at a time.. All the servers in the group are connected to the SQL Cluster / oracle RAC .. I still see no outages, vice a bomb or something physical/network related. DR is the fail over to a completely redundant system in a different location, where either data replication was going on, or the DSO was turned to active to a live system, which is the case of the bomb/tornado/911 instance. On Wed, Mar 19, 2014 at 1:06 AM, Joe D'Souza <[email protected]> wrote: Within certain limits though. I would not go that far to claim to the customer/management that there will be absolutely no down time during code migration. There will be. By taking servers on and off a server group, to upgrade core system versions, yes that can be done with 'minimal' down time. But the migration and code upgrade, takes as much down time as the migration of the code itself takes. Even if you stand up a completely new parallel system, and then decide a switch by mirroring a database, there still will be that minimal time required to port the delta data. Personally I think it is not possible to completely eliminate downtime if your system is significantly large. Its like approaching infinity in mathematics - you can get close, but you can never get there. You just got to be content you got close enough.. Cheers Joe -----Original Message----- From: Action Request System discussion list(ARSList) [mailto:[email protected]] On Behalf Of Zandi Sent: Tuesday, March 18, 2014 6:49 AM To: [email protected] Subject: Re: BMC Remedy and Flash 24/7 is already there... It is called server groups, if you implement this would can take a server down and the others will takeover while it is being patched. You will need a load balancer as well. This also allows for larger system use as well My 2 cents Sent from my iPhone > On Mar 17, 2014, at 3:12 PM, James Smith <[email protected]> wrote: > > Nice info Doug, thanks for sharing. Want to add 2 cents if its considered then its of great use. > > Currently we have windows based tools for development activities and data migration like Developer studio and Import tool. Will it be feasible to make then available over web? > > One more thing, how can we make remedy to be available 24*7 during upgrades as well - zero downtime upgrades. This will help the product to compete in the market. ____________________________________________________________________________ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org "Where the Answers Are, and have been for 20 years" -- Patrick Zandi _ARSlist: "Where the Answers Are" and have been for 20 years_ _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org "Where the Answers Are, and have been for 20 years"
