Wired Intelligent Edge

I have seen some posts about Esxi hosts and whatnot, but I am having an issue and I think I need to reconfigure the vSwitch from standard to distributed. 

I am upgrading from a 5900 to a VSX cluster. I have it all set up and most things migrated over to the VSX cluster already with no issues.

I am having trouble with the Esxi hosts though. The vSwitch is currently set up as standard. I know I need to change it to a distributed vSwitch in order to do trunking and LACP from what I have read. 

I migrated several cables over from the 5900 to the VSX cluster with a work around until I change the vSwitch.

The issue I am seeing is that when I move certain links over to the VSX cluster, the switch sees them as being up, it sees their transceivers and all. When I check them on the Esxi Hosts web GUI, the links say they are dead. 

I can't do trunking or use LACP because it is set up as a standard vSwitch if I am understanding correctly and I think that is where the issue is, but before I start bugging my Sys Admin for some help doing it, has anyone seen that before where the switch says the link is up, but the Esxi hosts web gui says the link is dead? Currently the work around I have been doing until I get the Distributed vSwitch is I just  configure the port as follows:

no shutdown
description Esxi host
no routing
vlan trunk native 1
vlan trunk allowed 1,100

------------------------------
rford1219
------------------------------

Hello, technically speaking - if Links Aggregation with LACP is not a strict requirement (and, speaking about VSX, it should because they - VSX LAGs and Links Aggregation with LACP provided by vDS - fit perfectly together) you could fallback to:

(1) Independent uplinks to both VSX Members (forgetting to use any Links Aggregation - Static or LACP - on ESXi side)
(2) Static Links Aggregation (Non Protocol) to VSX LAGs defined on VSX (using Links Aggregation in its basic form on ESXi side <- this doesn't require vDS since it works simply with vSS).

I'm inclined to prefer (2) BUT from the VSX perspective (VSX LAGs perspective) I've to admit aggregate links with LACP is far a better option (and VSX side you can easily verify the VSX LAGs - what are Multi-Chassis LAGs - statues...a thing that using Non Protocol Links Aggregation doesn't happen...I'm referring to the output of show lacp interfaces multi-chassis CLI command)...that's to say that also (1) is an option.

The config:

no shutdown
description Esxi host
no routing
vlan trunk native 1
vlan trunk allowed 1,100

means that that interface is administratively Up, is operating in Switching mode and acts as a interface carrying more VLANs so in "Trunk operating mode" (with VLAN 1 as native so Untagged and VLAN 100 as Tagged)...that's an interface typically used for VLAN-aware hosts capable of passing and accepting traffic untagged with one VLAN id and tagged with various other VLAN ids.

------------------------------
Davide Poletto
------------------------------