I currently have an ASP.NET web app running on a cluster in a data center talking to a SQL Server DB. I need to make this setup redundant with a hot-hot DB in another data center as well as a replicated web cluster.
What architectural options are available to setup this type of fail-over?
Sure, gotcha, I apologize, i'm a bit new here, this is an awesome site!
For the proposed solution I would recommend 3 tiers of redundancy in a multi-threaded setup.
Database- Continue to use SQL in a cluster configuration, 2 nodes with a DRBD RAID10 pair of storage systems to keep the data up and running. In the event of Datacenter failure (worst possible case) have a 3rd tier passive replica at another datacenter with transaction logs being shipped.
Web- 2 Web servers, local data stored on a seperate set of drives RAID1 or RAID10 depending on size and I/O. Replicate data via Robocopy. Added performance and requirement for 100%, incorporate dedicated load balancers in a pair. Additional replica web server at a seperate datacenter.
it depends on how current you want your data and how much data you will be transfering.
noneil, I think the point of the site is to encourage community problem solving. Lets have some open discussion so other people can benefit from having this question answered here.
I do a lot of work involving these kind of solutions and meeting high levels of redundancy and compliance. I can do you a solid and offer recommendations.
let me know if you want to chat, have a good one.