ESXi 5.1 Ping issues

[TroyB]

Ok, so let me give you some info about our setup first:

vCenter Server version 5.1

Our cluster consists of 6 ESXi hosts, 5 of which are ESXi 5.0 and we currently have 1 that is 5.1.

Each host has a total of 4 physical NIC connection (8 total but the others are just for redundancy).

Each NIC has it's own VLAN:

NIC1 = Service Console - 192.168.100.x

NIC2 = Distributed Switch Trunked

NIC3 = vMotion - 192.168.200.x

NIC4 - Fault Tolerance - 192.168.75.x

Default VMKernel Gateway is 192.168.100.x (our default gateway for the 100VLAN)

All 6 of our hosts are set up and look 100% identical, HOWEVER:

Our "Management" VLAN is 192.168.75.x so all of our monitoring equipment, etc. (HP Systems Insight Manager, etc.) is on that VLAN.

Anything on the 75 VLAN can not ping the hosts service console that's been moved over to 5.1 even though any machine on any VLAN OTHER than 75 I can ping the host just fine.

Here's a quick look at how my routes look:

VMkernel Routes:
Network          Netmask          Gateway          Interface
192.168.75.0     255.255.255.0    Local Subnet     vmk2
192.168.100.0    255.255.255.0    Local Subnet     vmk0
192.168.200.0    255.255.255.0    Local Subnet     vmk1
default          0.0.0.0          192.168.100.1    vmk0

Any thoughts on what I can do to resolve this?  I find it very odd that the setup looks 100% identical to my 5.0 hosts and they don't see this issue at all.

[Sreejesh_D]

Is 75 is a layer 2 or Layer 3 VLAN?

Does VLAN 75 hosts/vms can ping the service console IPs of 5.0 hosts?

[TroyB]

It's Layer 3 and the yes they can.

[Rubeck]

If you try to remove the vmk2 port as a test does the service console respond then?

/Rubeck

[john_lauro]

If you click properties on the NIC, and then click on where you have the configuration for vMotion and IP storage for the vlan, make sure it says Management Traffic Enabled, otherwise you will not be able to ping it even though it will be in the routing table.

[TroyB]

Rubeck that actually worked!

So what does that tell me I need to do so that I can have the Fault Tolerance network added back in (That's what vmk2 was used for)?

[Rubeck]

I would move FT to a different subnet....

When an echo packet is recieved by the vmk0 interface, what interface is then used for trying sending the reply? vmk0 or vkm2? It could be vmk2 as this is directly connected to the source network from where it originates..... but I honestly don't know for sure.

/Rubeck

[john_lauro]

That would be my guess (it wants to use vmk2), and why you have to have management enabled on both vmk0 and vmk2 network in this case.

[TroyB]

I guess I'll open a case on thise.  Just not sure why this was not an issue with 5.0 set up the exact same way but now and issue with 5.1.