ESXi 5.5: VM can't ping a physical host in a vLAN

Deathmage

Hey guys,

So I'm being thrown in a loop on this one.

Basically I have a server that I'm connecting to a vLan on a switch. The default gateway of the ESXi server is the IP address of one of the L3 switches, I setup a nested Port Group within a IP hashed vSwitch with a vlan of say 250 and assigned it to the server (VM), I setup the VM with an ip address, say 192.168.100.5 with a mask of 255.255.255.224 or a /27 with a default gateway IP address on this vlan of 192.168.100.1 on the L3 switch.

Since this server have three vlans, one for the vlan in question for printers, and the other two for backups and normal traffic I need to setup a static route inside of windows so the traffic knows what port to sent it out. So say the default gateway of the L3 switch is 192.168.1.1 and the IP address of the vlan that isn't work is 192.168.100.5 and the interface of the virtual nic is 20 I do the follow static route "route add 192.168.100.0 mask 255.255.255.224 192.168.100.1 metric 1 if 20 -p"....


Now when I do this something really strange happens, the L3 switch is performing OSPF and it has a Area 0 and Area 15 and routing is happening correctly. But what is strange is the VM in question on the vlan in question it can ping itself (means the virtual nic works), it can ping 127.0.0.1 (local), it can ping the L2 switch the EC from the ESXi host is connected too, and it can ping the IP address on L3 switch in vlan 250, as can it ping the ip addresses of all the other vlans and resolve DNS names but what it can't do it ping the physical printer server.

The physical print server has the same static routes, but flipped, but likewise the print server can't ping into the ESXi host where the VM in question resides but can ping every other device in the network and the L2/L3 address of the vlan in question....what's strange is the other vlans go over the same EC trunk to the ESxi hosts. The vlan in question was added to the EC trunk. it's really just weird. it's just two hosts that won't talk to each other.... I've tried VMXNET 3 and E1000 nics and nothing is passing the traffic...


for the death of me I can't figure out why it won't connect....

Anyone think of something I'm missing here, maybe I've been banging my head on the walk for so long I'm missing it.

Deathmage

last ditch effort, thinking out side of the box here, I'm updating the firmware on the physical print server nic card.

Deathmage

Bingo, troubleshooting takes a long time, it was the nic driver, updated it and it worked!

I was about to scream if that didn't work or start talking out-loud to myself

techfiend

I've never heard of someone having to update NIC firmware but I've had to fix a few issues with BIOS updates. So much that it's quite high on my list of things to check.

higherho

Quick question, Are you doing your VLAN tagging on the switch or on the vSwitch? It has to be one or the other not both. Are you doing all your tagging in the vSwitch and the physical port is the trunk?

jack84ie

good one to check as well is Make sure each Port has a Trunk port. and its Native

Deathmage

indeed. switching was setup correctly.