The other day we were doing some maintenance on a host and performed some vMotions. During this time one of the hosts lost connectivity for less than a minute. After looking deeper into it we saw that the Cisco switch that the host is connected to blocked two of the VLANs. The vMotion that aligned with this event was the vMotion of a NetScaler VPX device that has two NICs that happen to be on the two VLANs mentioned (4 + 225). The host (and all other hosts in this cluster) are configured with 8 physical NICs in a port channel (this one being port-channel42). The host that the device vMotioned from was also on the same switch (in a separate port channel). We are unable to understand how this event is possible being that the NetScaler and ESXi vSwitch do not participate in STP. The NetScaler is operating in Layer 3 mode and would not send or relay any traffic like this.
Below is an excerpt from the switch log:
Jan 25 20:17:21.809: %SPANTREE-2-RECV_PVID_ERR: Received BPDU with inconsistent peer vlan id 4 on Port-channel42 VLAN225.
Jan 25 20:17:21.809: %SPANTREE-2-BLOCK_PVID_PEER: Blocking Port-channel42 on VLAN0004. Inconsistent peer vlan.
Jan 25 20:17:21.809: %SPANTREE-2-BLOCK_PVID_LOCAL: Blocking Port-channel42 on VLAN0225. Inconsistent local vlan.
Jan 25 20:17:36.810: %SPANTREE-2-UNBLOCK_CONSIST_PORT: Unblocking Port-channel42 on VLAN0004. Port consistency restored.
Jan 25 20:17:36.810: %SPANTREE-2-UNBLOCK_CONSIST_PORT: Unblocking Port-channel42 on VLAN0225. Port consistency restored.
