Symptom

Existing tunnel flaps when applying config It seems the tunnel is flapping after applying a config to an existing tunnel that is up. Here is some info that might help: This is the tunnel state before applying config: 14:57:15.638: Name: nxosA_t1000 (tunnel-te1000) Destination: 192.168.0.4^M 14:57:15.638: Status:^M 14:57:15.648: Admin: up Oper: up Path: valid Signalling: connected^M 14:57:15.648: path option 10, type explicit lsp0 (Basis for Setup, path weight 80)^M 14:57:15.648: path option 20, type explicit lsp1^M 14:57:15.648: ^M 14:57:15.648: Config Parameters:^M 14:57:15.648: Bandwidth: 1000 kbps (Global) Priority: 7 7 Affinity: 0x0/0xffff^M 14:57:15.658: Metric Type: TE (interface)^M 14:57:15.658: AutoRoute: enabled LockDown: disabled ^M 14:57:15.658: auto-bw: disabled^M 14:57:15.658: Active Path Option Parameters:^M 14:57:15.658: State: explicit path option 10 is active^M 14:57:15.658: BandwidthOverride: disabled LockDown: disabled Verbatim: disabled^M 14:57:15.669: ^M 14:57:15.669: ^M 14:57:15.669: InLabel : - ^M 14:57:15.669: OutLabel : Ethernet2/1, 20^M 14:57:15.669: RSVP Signalling Info:^M 14:57:15.669: Src 192.168.0.1, Dst 192.168.0.4, Tun_Id 1000, Tun_Instance 5^M 14:57:15.669: RSVP Path Info:^M 14:57:15.669: My Address: 192.168.0.1 ^M 14:57:15.669: Explicit Route: 10.10.10.2 14.14.14.2 14.14.14.4 192.168.0.4 ^M 14:57:15.679: Record Route: NONE^M 14:57:15.679: Tspec: ave rate=1000 kbits, burst=1000 bytes, peak rate=1000 kbits^M 14:57:15.679: RSVP Resv Info:^M 14:57:15.679: Record Route: 10.10.10.2 14.14.14.4 ^M 14:57:15.679: Fspec: ave rate=1000 kbits, burst=1000 bytes, peak rate=1000 kbits^M 14:57:15.679: Shortest Unconstrained Path Info:^M 14:57:15.679: Path Weight: 80 (TE)^M 14:57:15.679: Explicit Route: 13.13.13.1 13.13.13.3 16.16.16.3 16.16.16.4 ^M 14:57:15.689: 192.168.0.4 ^M 14:57:15.689: History:^M 14:57:15.689: Tunnel:^M 14:57:15.689: Time since created: 1 minutes, 2 seconds^M 14:57:15.689: Time since path change: 11 seconds^M 14:57:15.689: Number of LSP IDs (Tun_Instances) used: 5^M 14:57:15.689: Current LSP:^M 14:57:15.689: Uptime: 14 seconds^M 14:57:15.689: Selection: reoptimization^M 14:57:15.689: Prior LSP:^M 14:57:15.689: ID: path option 20 [3]^M 14:57:15.699: Removal Trigger: reoptimization completed^M Applying config here (Same as the config that was already applied except added path-option 5): 14:57:16.052: ^MnxosA(config)# mpls traffic-eng configuration^M^M 14:57:16.395: ^MnxosA(config-te)# explicit-path name lsp0^M^M 14:57:16.739: ^MnxosA(config-te-expl-path)# index 10 next-address strict 10.10.10.2^M^M 14:57:17.082: ^MnxosA(config-te-expl-path)# index 20 next-address strict 14.14.14.4^M^M 14:57:17.455: ^MnxosA(config-te-expl-path)# mpls traffic-eng configuration^M^M 14:57:17.789: ^MnxosA(config-te)# explicit-path name lsp1^M^M 14:57:18.152: ^MnxosA(config-te-expl-path)# index 10 next-address strict 11.11.11.2^M^M 14:57:18.516: ^MnxosA(config-te-expl-path)# index 20 next-address strict 15.15.15.4^M^M 14:57:18.869: ^MnxosA(config-te-expl-path)# interface tunnel-te1000^M^M 14:57:19.232: ^MnxosA(config-if-te)# destination 192.168.0.4^M^M 14:57:19.616: ^MnxosA(config-if-te)# path-option 10 explicit name lsp0^M^M 14:57:19.969: ^MnxosA(config-if-te)# path-option 20 explicit name lsp1^M^M 14:57:20.333: ^MnxosA(config-if-te)# path-option 5 explicit name lsp1^M^M 14:57:20.676: ^MnxosA(config-if-te)# autoroute announce^M^M 14:57:21.020: ^MnxosA(config-if-te)# bandwidth 1000^M^M 14:57:21.383: ^MnxosA(config-if-te)# ip unnumbered Loopback0^M^M 14:57:21.747: ^MnxosA(config-if-te)# record-route^M^M 14:57:22.090: ^MnxosA(config-if-te)# path-selection metric te^M^M 14:57:22.453: ^MnxosA(config-if-te)# no shutdown^M^M 14:57:22.807: ^MnxosA(config-if-te)# end^M^M Checking state of tunnel here (observe tunnel is DOWN!): 14:57:25.947: Name: nxosA_t1000 (tunnel-te1000) Destination: 192.168.0.4^M 14:57:25.957: Status:^M 14:57:25.957: Admin: up Oper: down Path: not valid Signalling: Down^M 14:57:25.957: path option 5, type explicit lsp1^M 14:57:25.957: path option 10, type explicit lsp0^M 14:57:25.957: path option 20, type explicit lsp1^M 14:57:25.967: ^M 14:57:25.967: Config Parameters:^M 14:57:25.967: Bandwidth: 1000 kbps (Global) Priority: 7 7 Affinity: 0x0/0xffff^M 14:57:25.967: Metric Type: TE (interface)^M 14:57:25.967: AutoRoute: enabled LockDown: disabled ^M 14:57:25.967: auto-bw: disabled^M 14:57:25.967: ^M 14:57:25.967: Shortest Unconstrained Path Info:^M 14:57:25.967: Path Weight: 80 (TE)^M 14:57:25.977: Explicit Route: 13.13.13.1 13.13.13.3 16.16.16.3 16.16.16.4 ^M 14:57:25.977: 192.168.0.4 ^M 14:57:25.977: History:^M 14:57:25.977: Tunnel:^M 14:57:25.977: Time since created: 1 minutes, 13 seconds^M 14:57:25.977: Time since path change: 8 seconds^M 14:57:25.977: Number of LSP IDs (Tun_Instances) used: 29^M 14:57:25.987: Prior LSP:^M 14:57:25.987: ID: path option 10 [6]^M 14:57:25.987: Removal Trigger: path verification failed^M 14:57:25.987: Last Error: PCALC:: Can't use link 10.10.10.1 on node 192.168.0.1^M Tunnel comes back up after about 6-9 seconds after applying config The script tries reoptimizing at this point and is waiting for a reoptimizing message here but it never comes since the tunnel actually flapped to the new config and nothing to reoptimize to resulting in script failing the test case. It does not always flap, only certain number of runs result in the tunnel flapping. This is with titanium-gdb.7.3.0.TE.0.9.iso from hsk_rsvpte branch

Conditions

4 node square topology double linked

Workaround

Tunnel comes up after a few seconds

Further Problem Description