Clustering, Log shipping and Database mirroring

  • Can Clustering, Log shipping and Database mirroring can be simultaneously ?

    In my company we are changing our ERP from As400 to a Microsoft Product and we are going to setup a failover cluster. I have read in Micosoft technet that we can have 'Database Mirroring and Log Shipping' and also 'Database Mirroring and Failover Clustering' and I wonder if we can combine both of them. If not what do you suggest for a company with about 350 users for data availability.

    Thank you,

    John

  • The question really isn't can you do this, but what are you trying to accomplish? high availability? zero-data loss? Minimal data-loss?

    Clustering is there for high availability but does not guarantee zero data-loss.

    Mirroring can give you zero data-loss in the event of a failure but there are latency costs associated with a two-phase commit.

    Log-shipping can help with low data-loss..

    There are costs and trade-offs associated with all of these. You need to specify what you want, and what you are willing to pay for in terms of $$$, complexity, and performance.

    CEWII

  • Hi,

    First of all thank you for your answer. The aim is high availability and minimal data loss.

    From what I have read, I have come to the conclusion for clustering and mirroring instead og log shipping.

    thanks,

    John

  • When clustering and mirroring is combined together there are some things you need to take care of.

    Eg. DB Mirror failover time is quicker than that of failover clustering. During normal cluster failover, db mirror

    will assume principal db is down and the mirror db would become principal. Again, there are different configuration modes in mirroring. You may need to test those scenarios completely before implementing.

    M&M

  • I will admit up front that I would rarely suggest db mirroring. One of my biggest issues is that for reliability SQL has to do a two-phase commit, meaning it has to write the same command to both the local and remote servers and both have to apply it before the next statement is executed, all the while the client is waiting. I especially don't like it in relation to use with a cluster due to the failover time issues. However there are times when it makes sense..

    For HA and the acceptability of a little data-loss clustering and log shipping with tight windows is a great solution..

    CEWII

  • you can go with sql clustering + san clustering..

  • i am doing something similar and need some architecture docs. Can any of you point me in the right direction?

  • Elliott Whitlow (2/10/2011)


    I will admit up front that I would rarely suggest db mirroring. One of my biggest issues is that for reliability SQL has to do a two-phase commit, meaning it has to write the same command to both the local and remote servers and both have to apply it before the next statement is executed, all the while the client is waiting. I especially don't like it in relation to use with a cluster due to the failover time issues. However there are times when it makes sense..

    For HA and the acceptability of a little data-loss clustering and log shipping with tight windows is a great solution..

    CEWII

    I would go with Elliott - > Clustering + Log shipping is a good solution.

    M&M

Viewing 8 posts - 1 through 7 (of 7 total)

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