L2VPN with Standalone Edge and local egress

[uniwre]

Hello,

I configured L2VPN between my NSX envirionment(SERVER) and a standalone Edge at a customer site(CLIENT) to stretch a layer 2 network. Both Edges use the same ip address.

Local egress is enabled at the Standalone Edge site, not on my NSX Edge.

While this L2VPN is UP, some clients at the customer site randomly lose connectivity to the network that is stretched by L2VPN. I personally think this is an ARP issue.

This leads to my question regarding Local egress configuration on the L2VPN Server which is the NSX Edge.

Can somebody tell me what local egress at the L2VPN server exactly means?

I read different things on the internet saying it is only used for internet traffic so traffic does not have to travel de L2VPN tunnel multiple times.

I also read that enabling local egress at the L2VPN server ensures that ARP requests are filtered so the L2VPN does not reply to ARP request.

Thanks in advance!

[Sreec]

The purpose of enabling egress optimization is to ensure packets are routed locally by respective ESG ( For eg : Site A OR Site B edge ) Egress IP which remains local to the site. Other wise even a ARP request will even go via tunnel .  While doing a ARP check you can  perform below steps to know the behavior

1)  debug packet capture interface name on L2 VPN Server Edge and same step on L2 VPN Client Edge

Local egress is enabled at the Standalone Edge site, not on my NSX Edge.

You should do this on both the sides (Client and Server)