How MPLS LDP-IGP Synchronization Works
Packet loss can occur because the actions of the IGP and LDP are not synchronized. Packet loss can occur in the following situations:
•When an IGP adjacency is established, the router begins forwarding packets using the new adjacency before the LDP label exchange completes between the peers on that link.
•If an LDP session closes, the router continues to forward traffic using the link associated with the LDP peer rather than an alternate pathway with a fully synchronized LDP session.
The MPLS LDP-IGP Synchronization feature:
•Provides a means to synchronize LDP and IGPs to minimize MPLS packet loss.
•Enables you to globally enable LDP-IGP Synchronization on each interface associated with an IGP Open Shortest Path First (OSPF) or IS-IS process.
•Provides a means to disable LDP-IGP Synchronization on interfaces that you do not want enabled.
•Prevents MPLS packet loss due to synchronization conflicts.
•Works when LDP is enabled on interfaces using either the mpls ip or mpls ldp autoconfig command.
To enable LDP-IGP Synchronization on each interface that belongs to an OSPF or IS-IS process, enter the mpls ldp sync command. If you do not want some of the interfaces to have LDP-IGP Synchronization enabled, issue the no mpls ldp igp sync command on those interfaces.
If the LDP peer is reachable, the IGP waits indefinitely (by default) for synchronization to be achieved. To limit the length of time the IGP session must wait, enter the mpls ldp igp sync holddown command. If the LDP peer is not reachable, the IGP establishes the adjacency to enable the LDP session to be established.
When an IGP adjacency is established on a link but LDP-IGP Synchronization is not yet achieved or is lost, the IGP advertises the max-metric on that link.
MPLS LDP-IGP Synchronization Requires Peer To Be Reachable
When the MPLS LDP-IGP Synchronization feature is enabled on an interface, LDP determines if any peer connected by the interface is reachable by looking up the peer's transport address in routing table. If a routing entry (including longest match and/or default routing entry) for the peer exists, LDP assumes that LDP-IGP Synchronization is required for the interface and notifies the IGP to wait for LDP convergence.
This requires that the routing table be correct and accurate for peer's transport address. If the routing table shows there is a route for the peer's transport address, that route must be able to reach the peer's transport address. However, if the route is a summary route, default route, or a statically configured route, it might not the correct route for the peer. You must verify that the route in the routing table can reach the peer's transport address.
When the routing table has an inaccurate route for peer's transport address, LDP cannot set up a session with the peer, which causes the IGP wait for LDP convergence unnecessarily for the sync holddown time.
MPLS LDP-IGP Synchronization Compatibility with IGP Nonstop Forwarding
The MPLS LDP-IGP Synchronization feature is not supported during the startup period if IGP NSF is configured. The MPLS LDP-IGP Synchronization feature conflicts with IGP NSF when the IGP is performing NSF during startup. After the NSF startup is complete, the MPLS LDP-IGP Synchronization feature is supported.
MPLS LDP-IGP Synchronization Compatibility with LDP Graceful Restart
LDP Graceful Restart protects traffic when an LDP session is lost. If an interface that supports a Graceful Restart-enabled LDP session fails, MPLS LDP-IGP Synchronization is still achieved on the interface while it is protected by Graceful Restart. MPLS LDP-IGP Synchronization is eventually lost under the following circumstances:
•If LDP fails to restart before the LDP Graceful Restart reconnect timer expires.
•If an LDP session restarts through other interfaces, but the LDP session on the protected interface fails to recover when the LDP Graceful Restart recovery timer expires.
https://www.cisco.com/c/en/us/td/docs/ios/12_0s/feature/guide/fsldpsyn.html