Geo Replication

The Geo-replication is used in inter-rack communication and for POD or VIP or BFD monitoring within the rack. The Geographic Redundancy comprises with the following:

  • Two instances of Geo pods are running for each rack.

  • Two Geo pods functions in Active-Standby mode.

  • Each Geo pod instance is spawned on a different Proto node or VM.

  • Geo pod running on the Proto node or VM having VIP is Active Geo pod.

  • In the event of Active Geo pod restart, VIPs get switched to other Proto node or VM and Standby Geo pod running on the other Proto node/VM becomes active.

  • Geo pod uses host networking mode (similar to UDP-Proxy).

  • Geo pod uses two VIPs:

    • Internal: VIP for Inter-POD communication (within the rack)

    • External: VIP for Inter-rack Geo pod communication

      It configures only on Proto Nodes on the L2 Subnet. It's used to communicate across the racks. This node has external connectivity to other Rack.

  • Logical-NF-InstanceID must be configured same for both cnSGW-CsSMFs in GR-Pair.

  • For KeepAliveD monitoring:

    • Geo pod uses base port as: 15000+(Logical-NF-InstanceID * 32) + 4

      Geo pod base port must be different than BGP speaker pod port.

      • The default port (without logical cnSGW-CSMF) as: 15004

      • For Logical cnSGW-CSMF configured with logical-nf-instance-id as 1, and then the port as: 15036

    • UDP-Proxy pod uses base port as: 28000+Logical-NF-InstanceID.

      • The default port (without logical cnSGW-CSMF) as: 28000

      • For Logical cnSGW-CSMF configured with logical-nf-instance-id as 1, and then the port as: 28001

    • BGPSpeaker-pod uses default base port as: 20000+(Logical-NF-InstanceID * 32)+4.

      • The default port (without logical cnSGW-CSMF) as: 20004

      • For logically cnSGW-CSMF configured with logical-nf-instance-id as 1, and then the port as: 20036

Note

Only ETCD and cache pod data gets replicated to the standby rack.