Procurve Switches and Windows Network Load Balancing in Multicast Mode causing high collision and drop rates

,

Over the last few days we have been looking at getting a Client Access Array going for our Exchange 2010 setup for basic redundancy and load balancing. I thought I would outline an issue we discovered with using Windows Network Load Balancing in a HP switching environment.

The first is whether to use Unicast or Multicast modes for the NLB Traffic. It is important to remember that Exchange does not care if you use Unicast or Multicast and is entirely dependent on your switching environment. At first we were confused as to which to choose due to the myriad of documentation suggesting either protocol. So we decided to give multicast a go.

After a few minutes our core switch (an 8212zl) started dropping packets for our services VLAN. We shut off the NLB and everything had returned to normal.  After a bit of digging we found that ProCurve switches are not compatible with multicast mode NLB due to their inability to have static ARP tables (they can only cache). But this can be remedied somewhat by issuing the following command on some models of switches:

ProCurve(config)# ip arp-mcast-replies

The command is supported on HP ProCurve E8200zl series, E5400zl series, E3500yl series, E6600 series and E6200yl series switches support Microsoft Windows NLB in Multicast mode.  Simply ensure that these switches are running K.15.03.0007 or greater firmware.


5 responses to “Procurve Switches and Windows Network Load Balancing in Multicast Mode causing high collision and drop rates”

  1. fletch00 Avatar

    Hi,
    we are experiencing unexplained “High collision or drop rate” events on several Procurve switches 3-4 times/day.
    What does NLB traffic look like so I can isolate it with a packet capture?

    thanks!

    1. John Avatar
      John

      Hi Fletch. If you’re running Microsoft’s NLB then use something like wireshark or ethereal you will see it as MS-NLB (Multicast or Unicast depending on how you’ve set it up). You are probably best creating a seperate VLAN for your NLB Heartbeat traffic or even go as far as creating a seperate physical network.

  2. Jeff Avatar

    I know this is an old post but I’m hoping you can help out here. We are not running NLB, however, we are using KeepAliveD with VRRP which utilizes multicast. I noticed on our ProCurve 3500yl switches that we have a high number of Dropped TX. Is this a problem with all multicast devices? I know some procurves support VRRP between them.

    1. John Avatar
      John

      VRRP utilises multicast to check on the status of the master router but I doubt that this would be the cause of the Dropped TX packets. AFAIK the 3500yl doesn’t support Static ARP lists but this generally only affects MS NLB.

      Have you checked to see if there are any other devices multicasting on the same subnet?

  3. Conrad Avatar
    Conrad

    This article is helpful. It saved me hours of troubleshooting.

    Thanks a mil

Leave a Reply

More Posts

Getting Folder Sizes and number of items in a Mailbox for a particular user on Microsoft Exchange using PowerShell

Recently one of our high-end users was going over their mailbox limit. In helping them to cut down I like to let them know what folders are using up the most of their quota (generally it is their sent items folder, but sometimes not). Executing the below PowerShell command in an Exchange Administration Shell gave […]

Changing the recovery mode doesn’t shrink an SQL Database log file, how to shrink logs manually.

So I found out recently that one of our servers was running out of space.  It’s our AV server so I was like what the hell, why is it running out.  Turns out it had an instance of SQL Server on there as a quarantine and configuration database.  The Virtual Machine was being backed up […]

Thoroughly cleaning up a WSUS server

I was recently tasked with performing a clean-up of some of our servers, removing old files/software installations as well as a clean-up of our WSUS server.  After a quick look I could see that our previous administrator had set it to download Driver updates as well, which was taking up quite a large amount of […]