locked
Implementing High availability RRS feed

  • Question

  • I am implementing a solution for high availability on my production database servers. We have some options like:

    1) Have 2 nodes in a cluster ( one in America  and the other in Europe)

    2)Have 2 nodes in a cluster( both in the same city in Europe) and also use databse mirroring to replicate changes to another database instance in America.

    3) Have 2 nodes in a cluster( both in the same city in Europe) and also use log shipping to replicate changes to another database instance in America.

    I would like to know what is best in terms of performance and availability, I would go for the second option but still would like some opinions from the experts in this forum.

     

    Thursday, May 6, 2010 12:49 AM

Answers

  • You need to know and understand the latency between sites. The latency will affect your technology choice however looking at your options, because they are geographically dispersed which will most likely have a reasonably high latency,  I would recommend option 2 with asynchronous mirroring.

     Microsoft has a good test case which will give you a real world example http://msdn.microsoft.com/en-us/library/ee355221.aspx

     

    • Marked as answer by LuisGranados Friday, May 7, 2010 2:01 PM
    Friday, May 7, 2010 12:41 AM

All replies

  • You need to know and understand the latency between sites. The latency will affect your technology choice however looking at your options, because they are geographically dispersed which will most likely have a reasonably high latency,  I would recommend option 2 with asynchronous mirroring.

     Microsoft has a good test case which will give you a real world example http://msdn.microsoft.com/en-us/library/ee355221.aspx

     

    • Marked as answer by LuisGranados Friday, May 7, 2010 2:01 PM
    Friday, May 7, 2010 12:41 AM
  • Thank you very much, yes I thought the same thanks to confirm.
    Friday, May 7, 2010 2:02 PM