Configuring a DR site
Posted by AQORN-Adam, 06-29-2008, 12:31 AM Hi everyone, I'm a programmer and not exactly a web config guy so I have a question: Is there somewhere that either has a tutorial or explains how a DR site is setup and activated when the primary site goes down? And how do you configure it to fail-over to the primary site again once it is back up and running? I'm planning to have a particular hardware configuration in place at a primary site (load balancers/web/app/database/SAN) supporting a service that MUST remain online. Because it's mission critical, I also will need a DR (disaster recovery) configuration at a secondary site. My challenge is how to configure DNS or whatever to fail over the primary site to the secondary site if the primary experiences a failure of some sort. Then fail it back post-recovery. I am planning to use the secondary site to burn-in development prior to go-live and when ready, migrate the changes to the primary so both sites are identical. I am doing some investigation regarding NetScaler appliances to understand how they work. At first glance it seems they are able to direct incoming traffic to specific locations based on various criteria (geo, speed, load capacity, etc). I noticed they are also a recommended solution by Citrix for traffic management... but a DR solution is evading me. And something is telling me it's not all about some piece of hardware. Whatever the solution is, it needs to be rock solid. Suggestions/thoughts?
Posted by Coolraul, 06-29-2008, 12:41 AM Must it be instantaneous failover or can it be down for 5, 10 or 30 minutes. This makes a big difference. Are you planning to have the sites all on a network operated by you or in "public" datacenters with different ip space?
Posted by AQORN-Adam, 06-29-2008, 05:45 PM The Primary Site and the Secondary DR Site would of course be in different data centers. In all likelyhood, the IP space would be different but we aren't necessarily looking to purchase a block of IP's from anyone or investigated VPN tunneling between the two sites so it's hard to say how the IP structure would be similar or different. Perhaps an IP engineer would know. I do know however that all machines behind the firewall/router(s) should have private addresses. Ideally any switch between the two sites due to a failure or manually reverting back to the Primary Site should be totally unnoticable/ transparent to the end user. Last edited by AQORN-Adam; 06-29-2008 at 05:53 PM. Reason: Spelling