in

Richmond SQL Server Users Group

Richmond VA's community oriented SQL Server User Group
This Months Sponsor:Synigent

Comparing SQL Mirroring to Clustering

Last post 03-30-2008 6:11 PM by grayp. 2 replies.
Page 1 of 1 (3 items)
Sort Posts: Previous Next
  • 03-21-2008 6:33 AM

    • wkhazzard
    • Top 10 Contributor
    • Joined on 03-21-2008
    • Richmond, Virginia
    • Posts 2

    Comparing SQL Mirroring to Clustering

    I'd like to hear the thoughts of other DBAs and Architects who've used both SQL Mirroring and SQL Clustering. My company, SnagAJob.com, is currently using SQL Mirroring for our production OLTP databases. We have issues almost every day: mysterious failovers, ADO.NET clients that repeatedly attempt to connect to the mirrored server instead of the primary (because they believe it's failed over). We're fairly certain that it's not networking issues because we see these same issues in our lab environment when everything is connected on a single gigabit switch. This would be a great topic for a future meeting, in my opinion, especially as more of a round-table type of discussion.

    W.Kevin Hazzard
    http://www.gotnet.biz/Blog
  • 03-21-2008 7:52 AM In reply to

    • fibrock
    • Top 10 Contributor
    • Joined on 03-20-2008
    • Richmond VA
    • Posts 40

    Re: Comparing SQL Mirroring to Clustering

    I like the idea of a round table on this topic, because not many of us have a lot of experience with clustering and or mirroring.  Currently we use both Clustering and Mirroring for our production environments.  Clustering has given us more of a headache than Mirroring due to some SAN management issues with presentation of disk to the cluster and the ID's that the OS is assigning these disks.  We have worked these issues out with our SAN team and they are no longer an issue but we switched to Mirroring for a few of applications because of this.  We have not had any significant issues with mirroring, which is running on an OLTP database that controls a pricing system.  While this is an OLTP system it does have a low level of transactions.  We run without a witness and therefore without automatic fail over, in the high performance asynchronous mode.  The reason for this is so we do not get any false system failovers, and our applications cannot take advantage of the newer SQL Server OLEDB drivers to allow for automatic failover.I like the idea of a few round table meetings a year in general.

     

    Steve Fibich
    http://vsteamsystemcentral.com/cs21/blogs/steve_fibich/default.aspx
    http://www.sqlservercentral.com/Authors/Articles/Steve_Fibich/525287/
  • 03-30-2008 6:11 PM In reply to

    • grayp
    • Top 10 Contributor
    • Joined on 03-30-2008
    • Palmyra, VA
    • Posts 5

    Re: Comparing SQL Mirroring to Clustering

    We do all clustering and no mirroring at work.   This is for 2000 OLTP systems, 2005 OLTP and AS clusters. 

     Most of our issues tie back to the San which we don't have control of.   We have had some ongoing failures that seem to occur when we are running full text index processes on one of the 2000 clusters. 

    We are in an active/passive configuration everywhere and while it is probably the most costly from the hardware perspective, we have the most comfort in the case that something fails.      

     This could be a great meeting in the future.

    Gray

Page 1 of 1 (3 items)
©2006-2008 Richmond SQL Server Users Group; ©2008 Richmond User Groups Corporation
Powered by Community Server (Non-Commercial Edition), by Telligent Systems