SQL Server Performance

Peer to Peer Replication - Node Failure Recovery

Discussion in 'SQL Server 2008 Replication' started by inquisitive, Jan 27, 2011.

  1. inquisitive New Member

    I am not a database administrator, how ever i have been looking into peer to peer topolgy of microsoft sql server 2008 R2. I have a major doubt.
    all the msn maerial and blogs say - sql server peer to peer replication is a High Scalability + High Availibility strategy.
    the examples say that if i have 4 nodes in the peer to peer topology - eg Node1 London , Node 2 Berlin ,Node 3 California and Node 4 Tokyo.
    I understand that each of these nodes can handle writes and reads independently. so when i write into Tokyo all the records are replicated in the other 3 nodes -london , berlin and california.
    Nos my doubt is - How is failure handled in peer to peer topology. Theory say that if Tokyo is down the users at tokyo might still manage to use the node from berlin ,london or califoria and have live data. How exactly is this done?? Does this need to be configured in the application or is there some config in the sql server nodes?
    I am not sure if the sql connection string with parameter failover partner= berlin might be used in this scenario. can any one elaborate? thanks
  2. Luis Martin Moderator

    Check:
    http://sql-server-performance.com/Community/forums/t/32455.aspx
  3. satya Moderator

    Welcome to the forums.
    The p2p replication topology is built in and the application must have individual connection to the nodes, also the featuer is purely for scalability and nto for entire availablity, see http://technet.microsoft.com/en-us/library/ms151196.aspx fyi. For HA purpose you can look at failover clustering or database mirroring that suits the environment.

Share This Page