high availability

  • Hi Every one.

    I've got a solution to design, and not sure exaactly how to tackle this problem.

    we have 2 databases that "Speak" to each other permanently.

    The one database is the custem financial database that we own. It is only used locally.

    It hosts sales orders, purchase orders, stock, invoices, and of course the Ledger etc.

    Lets call it FinDB

    Then we have a database that connects to the dealer website. there retail sales are recorded and orders are placed etc.

    Lets call it RetDB

    Now RetDB uses a lot of views from FinDB to find stock allocated to dealers, and to place orders etc.

    Also FinDB reads retail sales data from RetDB, then combine with wholesale data for management reports and sales forcasting.

    Now:

    Our internet line went down for the 1st time in a year, and it seems like it will stay off for a large part of today.

    This of course prevents the dealers from using the dealer interface.

    Management, wants to move the RetDB to another network. THis is purely for higher availability.

    FinDB has some queries that will not be viable to run over a WAN, so it will have to stay on the local network.

    Which angle would be best to address this?

    Log shipping over a wan link? Replication?

  • LogShipping is only used as a stand-by-server in which case one would be primary and the other as a secondary server which is a stand by server where it does not accept any connections or transactions. All the log transactions would be transferred over the network to the stand-by-server from primary. I believe this does not meet your needs as you have to update data at both nodes.

    Take a look at Replication: Transactional Replication

    http://technet.microsoft.com/en-us/library/ms151176(SQL.90).aspx

    Lets wait and see other members views

Viewing 2 posts - 1 through 1 (of 1 total)

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