10 Responses

  1. Best Practices for Running SQL Server Virtualized — Long White Virtual Clouds | Farhan Parkar's Weblog

    […] via Best Practices for Running SQL Server Virtualized — Long White Virtual Clouds […]

  2. Justin Durrant
    Justin Durrant at |

    Great stuff as always! Thanks!

    Reply
  3. CARLOS ALFONSO
    CARLOS ALFONSO at |

    Hi Michael, great article!!. I have a questions. We have a VM monster SQL Machine on a nutanix block with 28 physical cores (2 x 14 and Hyperthreading enable) model 8035-g4, And since we installed Nutanix, I have assigned 2 physical processors with 12 cores each processor
    I have been doubt, which configuration is better, ratio 1:1 VCPU to pCPU i mean i must have assign 12 physical processor with 1 core each processor to obtain a better performance? or my configuratoin is ok?

    regards,

    Reply
  4. Rurik Wilmot
    Rurik Wilmot at |

    In your data file split image above…..you have the mount points listed as VMDK files……would these best practices need to be altered in any way if the mount points were actually iSCSI LUNS delivered via Nutanix ABS vs. the VMDK files?

    Reply
  5. Jesse Boyce
    Jesse Boyce at |

    Hi Michael,
    In the disk design image, with the 8 VMDKs being set as Mount Points on F:, where does F: itself live? I’m assuming it would be another VMDK that’s not on the diagram (unless I’m missing something obvious). Thanks.

    Reply
    1. @vcdxnz001
      @vcdxnz001 at |

      Hi Jesse, F: Drive is just another VMDK, usually very small, say 10GB, and the mount points just live on that.

      Reply
      1. Jesse Boyce
        Jesse Boyce at |

        Hi Michael,

        I thought it might’ve been something like that, thanks for the confirmation. I’m assuming it doesn’t matter which controller it’s on since there’s no real I/O going through it? (like, it could be on the first controller that has the OS VMDK)

      2. @vcdxnz001
        @vcdxnz001 at |

        Correct, doesn't matter which controller it goes to.

Leave a Reply