I've read the article, but I may be missing the actual cause of the problem. This is what we've been doing since our vCenter and ESX were at 4.x (we're on 5.1 now): - Create a Windows 2008 R2 prototype VM with a VMXNet 3 NIC and version 7 hardware. - After shutting down the prototype and before converting the VM to a template, remove the NIC. Note that we do *not* uninstall it from within the OS first. - Our provisioning process creates a VM from the template and adds a VMXNet 3 NIC. - We don't have any ghosts in Device Manager, the NIC name is "Local Area Connection" (as was the "previous" one) and we are not having any networking problems. We've deployed well over 100 of these over the last 3 years or so.
I'm just glad that we aren't seeing the problem and I appreciate your posts in case we do in the future, but again I'm not sure what's different about the way we're doing it and what's in the article. On Tue, Apr 8, 2014 at 1:49 PM, Kurt Buff <[email protected]> wrote: > And, the last bit of info that pretty much wraps this up: > > http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1020078 > > Kurt > > On Mon, Apr 7, 2014 at 5:40 PM, Kurt Buff <[email protected]> wrote: > > All, > > > > Well, after a bit more troubleshooting over the weekend, we did one last > test. > > > > We nuked the VMXNET 3 adapters on the VMs and replaced them with E1000 > > adapters, and now all is working. > > > > What I was told by cow-orkers is that the VMs on this host were copies > > of a template from an ESXi 5.1 host, and this host, as noted below, is > > running ESXi 5.5. > > > > Arguing against this being root cause is the fact that I did a clean > > install of two Win7 VMs with VMXNET adapters, and they showed the same > > problem communicating between themselves. > > > > I'm more suspicious of the fact that the 5.1 host is part of a vSphere > > Standard cluster, vs. the 5.5 host being the standalone free version > > (we'll be getting it up to Standard soonish, once the budgeting cycle > > is completed - I hope). > > > > At any rate, we're now proceeding with the migration of the mailboxes, > > etc., in that office. > > > > Kurt > > > > On Thu, Apr 3, 2014 at 1:39 PM, Kurt Buff <[email protected]> wrote: > >> All, > >> > >> My search-fu is failing, so I turn to you for help... > >> > >> I have a small ESXi 5.5 host, about to go into production. > >> > >> The three VMs (2008R2 for all of them, a DC, Exchange 2010 and a PRTG > >> box) on it can communicate with machines not on the ESXi host - ping, > >> RDP, etc. - and vice versa. No problems. > >> > >> However, the three VMs on this host cannot talk with each other. No > >> ping, no RDP. When pinging from one of the VMs to another, I get a mix > >> of unreachables from the VMs own address and straight timeouts. > >> > >> There is only one vSwitch, which has two NICs bound to it, and the > >> vswitch is set up to route based on IP hash. The physical switch to > >> which they are connect (and this shouldn't matter, but...) is an HP > >> 2510G-48, and the ports for the host are in a simple trunk - no LACP. > >> > >> I've turned off the Domain profile of the firewall on one of the > >> machine, which seems to make no difference. > >> > >> I've examined the VMware host security settings to no avail. I've > >> turned off the Windows firewall. > >> > >> I've got 3 ESXi hosts in a vSphere Standard cluster that doesn't have > >> this problem. > >> > >> Kurt > >> > >> > > >
