Weirdness....

[Jawloms]

I have 4 physical servers all running Hyper-V on Server 2016. Everything is fine except for 1 virtual machine on one host, and one virtual machine on another host. If I run a constant ping, they either respond in 1ms, or timeout. Nothing in between. Probably about 75% success rate. If I move the VM to a different host, the problem moves. It's only come on in the last couple of days. Both are running Server 2012 R2. One is a file server and one is my Sophos Enterprise server. Both can be pinged fine from their respective host with no packet loss. All hosts and all other VMs on the same hosts, and everything else on the network appears to be working as it should, it's just these two virtual machines. They haven't recently done Windows Updates. Pinging from the servers has the same results -> they can ping their host with no problem, but anything else results in random packet loss.

 

Answers on a postcard....?

 

Stuart

[mikeprice]

Drivers?? - is there an updated one somewhere???

[sted]

are they on teamed nics?

ive sometimes found that a specific vm on a team has issues move it to a dedicated card and the issue goes away. its infrequent but i have seen it and never got to the bottom of it

[strawberry]

recreate the nics and go from there.

[sister_annex]

Check the MAC addresses assigned to them... we've had this in the past where hosts give out a nic with a duplicate MAC

[Meldrew]

You've not managed to spin up 2 copies of the same VM on different hosts have you? (Been there, done that!)

[pantscat]

Check the MAC addresses assigned to them... we've had this in the past where hosts give out a nic with a duplicate MAC

 

This. Check this.

[Jawloms]

are they on teamed nics?

ive sometimes found that a specific vm on a team has issues move it to a dedicated card and the issue goes away. its infrequent but i have seen it and never got to the bottom of it

 

Looks like we have a winner! I've removed a NIC from the team, given it an IP, set up a virtual switch, assigned that to one of the VMs in question and all seems well now. Never had that before!

 

Thanks @sted, and to everyone else for your input.