Merge replication without network

  • Hi! I know that would sound crazy, but does anyone knows how to do a merge replication between two computers, hundreds of miles apart, but with no network connection (working in the woods, no satellite, etc., can't move the computer but USB key would be possible)?

  • tilew-948340 (9/16/2009)


    Hi! I know that would sound crazy, but does anyone knows how to do a merge replication between two computers, hundreds of miles apart, but with no network connection (working in the woods, no satellite, etc., can't move the computer but USB key would be possible)?

    Your right it does sound crazy, in theory it could be done but I would question why you would consider merge replication for doing this. doing a manual sync using the usb key would be the way you could do it in theory. but there would be better ways of keeping data in sync between two remote machines.

    --------------------------------------------------------------------------------------
    [highlight]Recommended Articles on How to help us help you and[/highlight]
    [highlight]solve commonly asked questions[/highlight]

    Forum Etiquette: How to post data/code on a forum to get the best help by Jeff Moden[/url]
    Managing Transaction Logs by Gail Shaw[/url]
    How to post Performance problems by Gail Shaw[/url]
    Help, my database is corrupt. Now what? by Gail Shaw[/url]

  • Well, I am all ear and open mind to any way that could do the job... Is a manual sync complicated? (note: I am new in the SQL world.)

  • I think if you can supply more details on what you are trying to achieve, it will be easier to give you an answer.

    Do you want the database on both computers to be the same, with changes made on both machines made to each other. and if so what would be the time difference you could allow between the machines being out of date with each other.

    --------------------------------------------------------------------------------------
    [highlight]Recommended Articles on How to help us help you and[/highlight]
    [highlight]solve commonly asked questions[/highlight]

    Forum Etiquette: How to post data/code on a forum to get the best help by Jeff Moden[/url]
    Managing Transaction Logs by Gail Shaw[/url]
    How to post Performance problems by Gail Shaw[/url]
    Help, my database is corrupt. Now what? by Gail Shaw[/url]

  • One guy would be in the wood doing environment data stuff (get, analyse, modify data)

    After one week, someone will replace him, continue his job and return when the first guy come back (so a USB key could be brought back in and out with each guy resulting in a sync every week).

    On the main SQL server, people would work on it doing analysis and report but also adding some data.

    On the main SQL server, there are table that the "wood guy" can't see.

    Being "up-to-date" on the data is the more efficient, so I tought that a merge replication would be the best way to do it, but as I said, there is no network cable in the wood...

    So the answers to your question are:

    yes, some tables of the data base would be the same, but not all table.

    yes, changes would be made on both machines (including "adding" process on both, so some index might result being the same?)

    One week of data changes between sync.

    Confusing, I know, isn't it?

  • Well, it's been more than 10 days since I ask the question. So, I guess that the merge replication is not possible in my scenario... :alien:

    I just have to write a program now to manage all that... I wish I could have more details about the "script" thing...

    Thank you all for your interest and time. It was worth asking anyway!

  • Here's a possible solution. Make a complete backup of the remote database and store that on USB key. Restore that to the main server when the guy gets to the office. Use ApexSQL's DataDiff product to sync up the data to the main tables. Note that you need some particular constructs and/or process controls to manage this if the home office people can add in new data that could have overlapping primary key values (such as identities).

    Best,
    Kevin G. Boles
    SQL Server Consultant
    SQL MVP 2007-2012
    TheSQLGuru on googles mail service

  • I was going to suggest the diff scenario. Apex Diff works well, as does Red Gate's Data compare (I work for Red Gate).

    I don't think merge would work unless the guy can bring the instance back, as in on a laptop. The logs need to be cleared, which means the publisher, or distributor, needs to see the subscriber.

    The other thing is that are the two guys coming back and going separately. Meaning that the wood guy comes back on Friday, you can work with data, and then the new wood guy goes back on Monday. That way you can determine changes.

    If they are changing in the woods, you'll never have a sync at that time. You'll have a week old db going out and trying to merge with one that's changed and is current. There can only be one true record of data. If that's how things work, I'd run the diff in the woods between two copies of the db.

  • ouch... "Apex Diff." I don't even know what your are talking about, but I'll check that. It sure would be easier if I knew more about SQL stuff, but it is the best way to learn, that's for sure!

    Thank you very, very, very much for your help.

  • Actually read my post carefully - it is ApexSQL DataDiff, a data comparison and synchronization application from ApexSQL. As Steve pointed out RedGate makes a competing product, and there are several others on the market as well IIRC.

    I will add for posterity's sake that data synchronization like you are attempting is just about the hardest thing you can face in the data world. Best of luck with your efforts!

    Best,
    Kevin G. Boles
    SQL Server Consultant
    SQL MVP 2007-2012
    TheSQLGuru on googles mail service

Viewing 10 posts - 1 through 9 (of 9 total)

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