With tunnel services enabled it’s time to actually configure the logical tunnel interfaces and bind the virtual routers together. This video walks through the complete configuration of the first inter-logical-system link — PE1 to PE2 — covering everything from interface addressing to peer unit binding and connectivity verification.

Each logical tunnel interface needs to be configured on both ends before it will function. On the PE1 side we configure LT000 unit 1 with an Ethernet encapsulation, an IP address from the 10.10.12.0/24 subnet, a description identifying the far end, and crucially a peer-unit statement pointing to unit 2 on the PE2 side. On the PE2 side we mirror this configuration — unit 2 pointing back to unit 1 as its peer. Without both peer unit statements configured the commit will fail.

Alongside the tunnel interface configuration we also set up the loopback interface and router ID for each logical system. These are needed later for routing protocols — configuring them now alongside the tunnel interfaces keeps everything in the right place from the start.

An important concept covered in this video is how Juniper treats the routing tables within logical systems. Unlike the global routing table which is shared across the whole device, each logical system has its own completely separate routing table. When you try to ping from the global context it won’t work — you need to either specify the logical system explicitly in the ping command, or switch your CLI context to the logical system using set CLI logical-system. We demonstrate both approaches and explain why the routing separation exists.

By the end of this video PE1 and PE2 are connected via a logical tunnel, both have loopbacks and router IDs configured, and we can successfully ping between them from within their respective logical system contexts. The same process is then repeated for the remaining PE connections in the topology.