For a lot of people it has been a crazy week. Some of you might wonder why, some of you know what I'm talking about. VMware Partner Exchange 2010. With PEX coming up for many of you that means GTJD, GTJD? Yeah, Getting The Job Done! Being away for a week in my case means I need to wrap up project and answer a lot of emails before it gets out of control. That doesn't however mean that I don't have time to create a top-5…. This weeks list contains the all-star bloggers:

  • Scott Drummonds – PVSCSI and Low-IO Workloads
    At low IOPS the CPU is doing very little work to access storage
    hardware.  In these environments it is simply not worth anyone’s time
    to implement and use a special driver storage driver.  But when 10-50k
    IOPS are streaming through the virtual SCSI bus, a new approach that
    halves the number of cycles spent on each IO will noticeably decrease
    CPU utilization.  This is why we created PVSCSI. The current design of PVSCSI coalesces based on OIOs only, and not
    throughput.  This means that when the virtual machine is requesting a
    lot of IO but the storage is not delivering, the PVSCSI driver is
    coalescing interrupts.  But without the storage supplying a steady
    stream of IOs there are no interrupts to coalesce.  The result is a
    slight increase to latency with little or no efficiency gain.
  • Frank Denneman – Sizing VMs and NUMA nodes
    ESX is NUMA aware and will use the NUMA CPU scheduler when detecting a
    NUMA system. On non-NUMA systems the ESX CPU scheduler spreads load
    across all sockets in a round robin manner. This approach improves
    performance by utilizing as much as cache as possible. When using a
    vSMP virtual machine in a non-NUMA system, each vCPU is scheduled on a
    separate socket.
  • Jason Boche – Configure VMware ESX(i) Round Robin on EMC Storage
    The answer was buried on page 88.  The nmp roundrobin setting useANO is
    configured by default to 0 which means unoptimized paths reported by
    the array will not be included in Round Robin path selection unless
    optimized paths become unavailable.  Remember I said early on that
    unoptimized and optimized paths reported by the array would be a key
    piece of information.  We can see this in action by looking at the
    device list above.  The very last line shows working paths, and only
    one path is listed for Round Robin use – the optimized path reported by
    the array.
  • Scott Lowe – Using IP-Based Storage with VMware vSphere on Cisco UCS
    From the VMware side of the house, since you’re using 10GbE end-to-end,
    it’s very unlikely that you’ll need to worry about bandwidth; that
    eliminates any concerns over multiple VMkernel ports on multiple
    subnets or using multiple NFS targets so as to be able to use link
    aggregation. (I’m not entirely sure you could use link aggregation with
    the 6100XP interconnects anyway. Anyone?) However, since you are
    talking Cisco UCS you’ll have only two 10GbE connections (unless you’re
    using the full width blade, which is unlikely).
  • Gabe – Licensing problems with VMware VIEW4
    The problem Jon was facing, was that it was impossible to just add
    those 20 (2×10) licenses to vCenter without assigning them to a host.
    Because, in our believe, there should just be 20 licenses in some sort
    of pool that each VDI VM would take one license from. It is possible to
    assign multiple hosts to one license so they can share the number of
    available VMs in that license. What you can’t do is have a host connect
    to more than one license, which in our opinion would also be feasible.