Routing Table flusing after a clear control connections

[maverick0]

I am looking for some Cisco documentation to clarify a behavior I am noticing on the cEdge router. When I perform a clear sdwan control-connections command on the router, the RIB is rebuilt. This router is also running other routing protocols in the service VPN (OSPF and BGP), but every time the control connection is cleared or the transport experiences any issues, the last update timestamp from the routes are refreshed, as seen in the debug ip routing output.

[MHM Cisco World]

sorry what is your Q, can you more elaborate 

thanks 

MHM

[maverick0]

Hi @MHM Cisco World 

Sure, let me try to provide more details.

I have noticed that when the control connection goes down and comes back up, or when I manually perform a 'clear sdwan control-connections' command, the routes in the RIB are flushed and rebuilt.

Below is a sample of the debug ip routing output for your reference.

*Dec 18 10:47:47.392 BRST: RT(1): rib update return code: 17
*Dec 18 10:47:47.392 BRST: RT(1): updating ospf X.X.X.X/32 (0x2) [local lbl/ctx:1048577/0x0] omp-tag:0 :
via 10.1.100.10 Gi0/0/3.1509 0 0 0x0 1048578 0x100001

*Dec 18 10:47:47.392 BRST: RT(1): rib update return code: 17
*Dec 18 10:47:47.392 BRST: RT(1): updating ospf Z.Z.Z.Z/32 (0x2) [local lbl/ctx:1048577/0x0] omp-tag:0 :
via 10.1.100.10 Gi0/0/3.1509 0 0 0x0 1048578 0x100001

*Dec 18 10:47:47.393 BRST: RT(1): closer admin distance for Y.Y.Y.Y, flushing 1 routes
*Dec 18 10:47:47.393 BRST: RT(1): add Y.Y.Y.Y/23 via 10.1.1.14, ospf metric [110/20]
*Dec 18 10:47:47.393 BRST: RT(1): updating ospf Y.Y.Y.Y/30 (0x2) [local lbl/ctx:1048577/0x0] omp-tag:0 :
via 10.1.1.18 Gi0/0/3.1509 0 0 0x0 1048578 0x100001

Thanks!

[MHM Cisco World]

I send you PM 

can you reply 

MHM

[Kanan Huseynli]

Hi,

I believe the reason is OMP restart.

Clear connections > resets DTLS/TLS between router and vSmart > OMP peering goes down > Routes which are cached due to graceful restart are used (as stale) > vSmart becomes available > Router flushes local cache, uses new OMP information over new OMP peering > due to local flush , redistributed routes are withdrawn in service side routing (OSPF/BGP etc.) > new routes are redistributed