Forum Replies Created

Viewing 15 posts - 76 through 90 (of 231 total)

  • RE: Are There That Many GUIDs?

    Craig-315134 (10/12/2010)


    Robert Frasca wrote:

    Can someone provide a legitimate business reason for using a GUID?

    UUIDs (universally unique identifiers) are valuable in distributed computing environments. Any one computing device may create...

  • RE: Are There That Many GUIDs?

    Other than for merge and updatable transactional replication, I've yet to hear an argument as to why an identifier has to be globallyunique. Sure, it sounds grand but could...

  • RE: Are There That Many GUIDs?

    Andy Warren (10/12/2010)


    ...Strangely perhaps, I believe that performance isn't everything. It's important, but sometimes I think we try too hard, the equivalent of taking out the seats in our car...

  • RE: Are There That Many GUIDs?

    André Lozeau (10/12/2010)


    GUIDs have the advantage that you can generate them before (or without) doing the insert on the table, which can be quite useful, especially in parent/child screens

    Yes they...

  • RE: Are There That Many GUIDs?

    It's been my experience that GUID's are frequently implemented incorrectly, for example, using the NEWID function instead of the NEWSEQUENTIALID function. This typically causes massive amounts of page splitting...

  • RE: Solid State drive for tempdb

    I would definitely make sure the "lock pages in memory" right is granted. That is a strongly recommended best practice for 64-bit SQL Server.

    It does sound like you might...

  • RE: Solid State drive for tempdb

    george sibbald (10/8/2010)


    The server itself has 38Gb memory assigned to SQL (64bit)

    Is that server supporting any other apps besides SQL Server? How much memory is on the machine total?...

  • RE: Solid State drive for tempdb

    It's kind of bizarre that you're not seeing any improvement. One of my colleagues at this client site told me they had moved TempDB to an SSD and saw...

  • RE: Solid State drive for tempdb

    One other thought on your test. It could be that your ETL processes aren't using TempDB in any significant way so you might not see a gain. Easy...

  • RE: Solid State drive for tempdb

    I found your test results very interesting. Thanks for posting. My guess is that the controller for the SSD isn't any faster than the controller for a "regular"...

  • RE: Solid State drive for tempdb

    getoffmyfoot (10/1/2010)


    getoffmyfoot (10/1/2010)


    Robert Frasca (10/1/2010)


    getoffmyfoot (10/1/2010)


    I've been dying to try this myself... something to keep in mind though, analyze the io stats of your db files first (see sys.dm_io_virtual_file_stats)
  • RE: Solid State drive for tempdb

    getoffmyfoot (10/1/2010)


    Robert Frasca (10/1/2010)


    getoffmyfoot (10/1/2010)


    I've been dying to try this myself... something to keep in mind though, analyze the io stats of your db files first (see sys.dm_io_virtual_file_stats)
  • RE: Solid State drive for tempdb

    getoffmyfoot (10/1/2010)


    I've been dying to try this myself... something to keep in mind though, analyze the io stats of your db files first (see sys.dm_io_virtual_file_stats) http://msdn.microsoft.com/en-us/library/ms190326.aspx .

    It could...

  • RE: Design

    UMG Developer (9/24/2010)


    Robert Frasca (9/24/2010)


    I'm confused as to why you don't think a field with the IDENTITY attribute is not a surrogate key or identifier. The only reason to...

  • RE: Design

    UMG Developer (9/24/2010)


    sknox (9/24/2010)


    So that gives us 2 identifier columns, remarkably similar to the two in this QotD -- and that's where you may have a point. But I'd caution...

Viewing 15 posts - 76 through 90 (of 231 total)