Post reply

Writing to the Passive Node

  • September 28, 2011 at 2:14 pm

    #1388068

    Thanks guys, I think those last two posts contain the info I was looking for.

    FWIW, I see that you can do multi-site clustering without the SAN, I'm just not sure the powers-that-be would be willing to bring in a third-party solution to handle the replication.

  • September 28, 2011 at 2:18 pm

    #1388074

    sickpup (9/28/2011)

    FWIW, I see that you can do multi-site clustering without the SAN,

    Yes, you can, but your inactive node is still inactive. You cannot write to both databases in any form of SQL HA (clustering, mirroring or log shipping).

    Gail Shaw
    Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
    SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability

    We walk in the dark places no others will enter
    We stand on the bridge and no one may pass

  • September 28, 2011 at 3:11 pm

    #1388120

    sickpup (9/28/2011)

    Thanks guys, I think those last two posts contain the info I was looking for.

    FWIW, I see that you can do multi-site clustering without the SAN, I'm just not sure the powers-that-be would be willing to bring in a third-party solution to handle the replication.

    Multi-site clustering certainly is possible but it is something that almost nobody does due to the costs. It would be cool to do, but I understand that Denali handles it much better but still is painful.. I like the idea of taking the SAN out of the equation but the costs....

    CEWII

  • September 30, 2011 at 1:43 am

    #1388709

    Clustering requires shared data storage between Active & Passive Nodes. Few softwares allow to share the local disk (am not talking about SAN) with multiple nodes but you would be badly hit on performance.

    I agree with the suggestion that you should either rely on SQL Server's clustering capability or completely rely on organization's clustering application.

  • September 30, 2011 at 1:59 am

    #1388716

    devendra_shirbad (9/30/2011)

    Clustering requires shared data storage

    For traditional cluster setups, yes.

    Geo graphically dispersed clusters negate this requirement. They remove the single point of failure on the storage and as such do not rely on single shared storage.

    However, to synchronously replicate your storage between sites requires extremely fast site links, not to mention purchasing the storage in the first place :w00t:

    You'll generally need multiple site links as your storage replication should be segregated for true performance and security. The costs are very high and you have to justify whether the outlays would benefit the organisation.

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • September 30, 2011 at 2:14 am

    #1388725

    I don't think we are talking about geo-clusters here.

Viewing 6 posts - 16 through 20 (of 20 total)

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