Merge Replication for HA / Failover

  • Hello All,

    We have 3 data centres, hosted in different locations - Lets call them A, B, C.

    A = Main database server

    B = Production Host (Web and DB Read only)

    C = DR (Replica of B)

    Currently, A replicates to B and C. B and C are read only databases. Our environment now needs another database which is transactional and hosted in Location B. This database needs to be available in location C for DR (replicated from B)

    At the moment, if we want to switch to DR, we change some DNS entries and the sites switch over to the other location i.e. C and nothing needs to be done from a database perspective as they are replicated from a master and the subscribers are read only. However! - Now that we are introducing a second transactional database in Location B, I need to put something in place so that if we flip from Prod to DR the data is already in location C and if updated there, the data will go to location B. Make sense 😛 !

    I was considering database mirroring, but that would mean a step to flip over databases (appreciate its just one more step). I was thinking of Merge replication so that Location B can write to Location C and vice versa without any intervention from a db perspective during failover. Generally speaking only 1 location would be active at anyone time.

    I wanted to get some thoughts on whether merge replication can be used for HA. Am I missing something in terms if a site goes down completely (i.e Production - B) and we flip to DR and need to rebuild B. As merge is bi-directional, does the publisher have to be setup from particular location?

    Any thoughts, advice, tips are greatly appreciated.

    Thanks

  • ****** Anyone? *******

  • I would take flipping a hundred switches to fail-over a mirror than deal with merge replication issues.

    Cheers

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply