performance of join across linked server

  • Dev (4/24/2012)


    Lynn Pettis (4/24/2012)


    I would look at Paul Randals blog on SQLSkills.com if you are looking for something more authoritative.

    The following still holds good...

    realyl need to find something from MS saying to do this... cant say "some well respected posters on SQLServerCentral SQLSkills said so"... I mean, that's good enough for me, but...

    Not sure what to take from this?

    I mean, Paul Randal was a project manager for Microsoft on the SQL Server product and wrote most of the code for SQL Server 2005 DBCC CheckDB. If he isn't an authoritative source, not sure who would be.

  • I am not arguing on Paul’s capabilities (and your suggestion as well). I just wanted to say that OP is particularly looking for MS reference.

  • Dev (4/24/2012)


    I am not arguing on Paul’s capabilities (and your suggestion as well). I just wanted to say that OP is particularly looking for MS reference.

    That's why I said I wasn't sure how to take your comment.

  • Dev (4/24/2012)


    I am not arguing on Paul’s capabilities (and your suggestion as well). I just wanted to say that OP is particularly looking for MS reference.

    Unlike Lynn, I DO know how to take your comments: Dude (or Dudette) there are a metric butt-ton of BAD MICROSOFT RECOMMENDATIONS OUT THERE, including "best practice" ones! Your insistence that MS must speak to a particular best practice is going to/has already lead you to trouble. I will (and you should) take advice given by Paul Randal on SQL Server over ANYTHING else you read to the contrary, or don't find at all, from other Microsoft sources. I could give you a list of others here on SSC for which that applies, but expect I won't get through anyway so I won't bother.

    Best of luck resolving your problem.

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

  • Well worded Kevin !

    RegardsRudy KomacsarSenior Database Administrator"Ave Caesar! - Morituri te salutamus."

  • Right- yeah, I dont need it to come from Microsoft, I just need it to be backed up some supporting data or documentation to understand why.

    All I meant is that if it's well documented "known fact" and there is an "official" article out there which has already done the metrics and testing and shows that yeah, you need to do that, then I dont really have to do that testing and write that doc myself in order to justify this to my peers and to my managment. If it's 'yeah everybody know that' that's one thing, but it doesnt seem that way to me.

    (I just had a quick look at Paul's blog and I dont see anythign about doing update stats fullscan after and ugprade from 2000- I could just be missing it)

    But beyond that- there IS some credence you have to give when it's a MS best practice- right or wrong. No one can fault a DBA for saying "look I did the research, reviewed the doc and followed the best practice recommendation from Microsoft". On the other hand, if I say, well, I couldnt find the official word from MS about this, but i had a really good discussion in a forum about this and the community seemed in agreement that this was the right thing to do... well you can see where I'm coming from.

    I mean, in reality, I am gonig to do the testing and I do value this discussion, but the fact of the matter is when Microsoft has something down as an "official" recommendation, there is a little bit more weight to that.

  • >> No one can fault a DBA for saying "look I did the research, reviewed the doc and followed the best practice recommendation from Microsoft".

    I absolutely can, and do on a regular basis. Honestly I am GLAD there are so many "worst best practices" out there, and so many bad defaults (1MB data file growth anyone), and design practices and maintenance practices and UDFs and ORMs-hitting-databases and, and, and...

    All of those things make for plenty of work for me!! 😎

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

  • Well, i hopefully you can explain why the best practice doesnt make sense to your clients, because so far all i've heard in this case is "microsoft isnt always right". Ok, yep. I agree. Now tell me why they are wrong in this case, or point me to something that says "this is why it makes sense to do X.".

    You cant just say " do x, i think its in a blog somewhere" and expect me to accept that as fact.

  • Following an upgrade from SQL Server 2000 you also need to run

    DBCC UPDATEUSAGE on all upgraded databases.

    For reference look at Best Practices on:

    http://technet.microsoft.com/en-us/library/ms188414(v=sql.105).aspx

    This may not be related directly to your issue but does stop some strange issues.

Viewing 9 posts - 16 through 23 (of 23 total)

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