1

I am trying to understand why this doesn't seem to work as I am thinking it should. I am in the middle of implementing VLAN in order to seclude different traffics such as voice, server, workstations... and I am still trying to understand how the un/tagging works

Here is the setup:

  • ESX Server on Dell SW with default VLAN 1 - so everything is untagged in VLAN 1 I have an LACP using p17-18

  • HP Procurve Server SW with default VLAN 1 - again everything is untagged in VLAN 1 I also have an LACP. the LACP status is active and good. the LACP trunk is in VLAN 1. I have a trunk to the Core switch below aka Trk1, this Trk1 is in VLAN 1.

  • HP procurve Core SW with default VLAN 1 - all is untagged in VLAN 1 I have a trunk to the server switch above aka Trk1, this Trk1 is in VLAN1. VLAN has a DHCP, through IP helpers.

My issue: Setting up a vSwitch in ESX not using tagging, everything works fine. Seeting up a vSwitch in ESX using tagging (here VLAN1), I loss connectivity to/from that VM.

I imagined that the explicit tagging of that virtual switch would not impact anything, but somehow it does and I do not understand why.

Did I miss something obvious?

2
  • Let me draw what I have VM <---> vSW <---all ports untagged on vlan1> Dell SW <lacp trk(vlan1)---lacp trk(vlan1)> Server SW <trk1(vlan1)---trk1(vlan1)> Core SW
    – KitkatNeko
    Dec 1, 2014 at 17:31
  • It's not entirely clear to me which link you are tagging. If it's VM <-tagged-> vSW <---all ports untagged on vlan1> .... you need to configure tagging on the VM interface and the vSW, just as if it was a physical machine and a real switch.
    – richardb
    Dec 2, 2014 at 9:25

3 Answers 3

1

this behaviour is because you begin to tag frames from the vSwitch and the Switch HP procurve Core SW is not tagging VLAN 1, this is usual because VLAN 1 is most frequently the default VLAN or Native VLAN. In most environments this VLAN is not tagging.

To make this work right:

1)you must untagged VLAN 1 in the vSwitch, and tag all the others VLAN that you want to create.

2) other option: you can tag VLAN 1 in your vSwitch and in the HP procurve Core SW.

I would prefer first option, but its matter of taste.

I Hope Helps!

1

If you draw graph of your network, you apply tagging on the links between nodes. You always need to apply tagging on the two systems connected to that link at the same time.

So, for a virtual switch, depending on which side you are tagging, you either need to tag the physical connection on the uplink port, or you need to change the VM to accept tagged packets. More likely the former.

0

More info might be better to help you out. But just based on what you described, it seems like this is not a LACP issue but VLAN trunking issue.

If you add VLAN into vswitch, vswitch then stops, then maybe the issue is on the vmware side. I assume you are using the standard vswitch, right? How did you configured your portgroup (VLAN or VLAN Trunking)?

Do you use a chassis based server/host? If so, how do you configured your chassis switch?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.