What is the "ideal" hardware configuration for SQL Server 2008

  • Thanks Greg. It great to find these predesigned solutions.

    One question though. Is the "Fast Track" optimized for only data warehouses? We do have a lot of OLTP too.

  • Fast Track is only for data warehouses. OLTP loads vary so widely.

    It's still hard to guess your load. Maybe Lynn is closed on memory, but 1GB a day/150 databases, isn't a lot. That's like 7MB a day/db, correct? Not a lot of writes.

    Depending on the batches, transactions, etc., you could have lots of variation on what you need. I'd also consider DR here, since I'm assuming having 150 dbs crash would be bad. I might even lean towards a 3 or 4 node cluster if you can do that, or even 3-4 servers (2 primary, 1-2 backup) with mirroring.

  • With the considerations for DR, one must also include considerations for Business Continuity.

    200K may seem like a lot of money but it can disappear in a heartbeat with the purchase of a SAN (depending on size), Backup Solution, DR Solution, and Business Continuity Solution. Ideally, you would want something that would not need replaced after five years.

    Jason...AKA CirqueDeSQLeil
    _______________________________________________
    I have given a name to my pain...MCM SQL Server, MVP
    SQL RNNR
    Posting Performance Based Questions - Gail Shaw[/url]
    Learn Extended Events

  • bdavey (5/4/2010)


    Thanks Greg. It great to find these predesigned solutions.

    One question though. Is the "Fast Track" optimized for only data warehouses? We do have a lot of OLTP too.

    Steve is correct - these are designed for huge data loads / query activity.

    But you may want to ask Microsoft more directly on specifics.

    We deal with a data warehouse, so it looks like something we will get to.

    But it does nothing for our AS load.

    Greg E

  • bdavey (5/3/2010)


    Thanks Lynn, Greg, & Steve. I think I see the error of may ways now. Let me define this a little more.

    It is a single cluster environment (1 active, 1 passive).

    It is about 90% OLTP (10% warehouse).

    It has 2 instances with 150+ databases.

    The total space consumed, excluding backups, is about 500 Gig.

    Space is expected to grow at about a 500 Mb/day.

    We expect to replace our hardware every 5 years.

    Rack mount is definitely prefered.

    Not concerned with SLA's at this point.

    Backup replication & storage does not need to be considered in this purchase.

    The majority of the growth occurs during normal U.S. business hours.

    $200k is the very maximum. I would probably scale down from this.

    Your are concerned with SLA or you wouldn't have a cluster.

    2 Dell R910 with 256GB of RAM each (go with the lowest number of procs) add more later if needed

    3 years of SAN space(reassess expected growth over time) depending on your infrastructure this could blow your budget or not be a factor.

    Run raid 10 for all.

    1TB of for the data lun

    250GB for the Log lun

    10 GB for SQL system (with MSDB moved to data and log luns)

    20 GB for SQL TempDB

    1GB for MSDTC LUN

    1 Idera SQL Diagnostic Manager license

Viewing 5 posts - 16 through 19 (of 19 total)

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