One of the problem faced in MPLS networks with LDP and IGP is when both are not synchronized.When we say synchronize we mean that both LDP and IGP should agree on an outgoing interface for a particular prefix.If by any means the LDP neighbor ship goes down but IGP remains UP, IGP will continue to use the link to forward the traffic which might get black-holed in MPLS VPN networks.When LDP breaks the IGP will forward the packet as IPV4 rather than Labelled packet, hence in MPLS VPN networks if the packet becomes unlabeled on any P/Backbone router it will get dropped as customer routes aren’t visible on P routers.
Another problem due to synchronization is that IGP converges much faster than LDP and hence begins to forward the traffic which might get dropped.A solution to above problem is the use of MPLS LDP-IGP SYNC.
COMMAND TO ENABLE LDP-IGP SYNC:
R1#router ospf 1R1(Config-if)#mpls ldp sync
This feature ensures link isn’t used to forward unlabeled traffic when the LDP session is down.Once the LDP comes up the link is used to forward the traffic then.
Let us use the simple topology below to understand how LDP-IGP SYNC works:
When LDP session is down, OSPF on R1 announces the link metric as maximum 65535 until synchronization is achieved.
After LDP session is established & label bindings are exchanged, the IGP advertises the link with its normal metric.