Four NE40Es constitute the core VPLS network. As the PEs, the NE40Es carry the VPLS service. The VLANs of the VPLS are attached to the NE40Es and bound to the VSI instance through sub-interfaces. One day, some users reported packet loss when they, at the subnodes of a VSI, tried to ping the central node.
1. Checking the central node, the engineer found that the NE40Es were connected with the central node through Lay 1 and Layer 3 interfaces, between which packets were transmitted normally. Thus, there must be nothing wrong with the central node or the links.
2. Checking with the customer, the engineer found that only users in one VSI were affected and no other services were affected. Thus, there must be nothing wrong with the NE40Es.
3. According to the MAC address of the central node reported by the customer, the engineer ran the disp mac-add xx-xx-xx command on the switch connected to the central node to check the MAC address. As a result, the engineer found that the MAC address of the central node flapped. Finally, under the flapping interface, the loop was found with an interface of the connected switch. The problem was solved after this interface was shut down.
1. Problem with the central node or the links.
2. Problem with the ARP entries in the VSI.
3. Problem with the NE40E boards.
As for the VPLS, we can see only Layer 2 entries on the devices. If all users within a VSI cannot successfully ping one another, generally, the problem is with Layer 3 entries as a result of the ARP virus or loop. For the problem caused by the loop, you can check whether a certain MAC address flaps. To check whether the problem is caused by the ARP virus, you can capture packets to verify your guess.