Cisco Bug: CSCvu80471 - VXLAN vPC VTEP - Extended traffic loss when vPC peer reloads before NVE source hold timer expiry
Oct 14, 2020
- Cisco Nexus 9000 Series Switches
Known Affected Releases
Under vPC VTEP (Fabric Peering and Physical), if user reload the one of the peer device and then reload the other one before the hold-down timer expired, the NVE interface will goes down. Symptom: Traffic loss for hosts behind vPC in a VXLAN setup. NVE interface remains down after source hold-down timer expiry: # show nve interface nve1 det <> Source Interface hold-down-time: 180 >>> default time Source Interface hold-up-time: 30 Remaining hold-down time: 67 seconds. <<< after countdown finishes, goes to 0 From the same device, we can see the following in the log: %ETHPORT-5-IF_DOWN_NONE: Interface port-channel1 is down (None) <<< Virtual peer-link goes down %USER-2-SYSTEM_MSG: NVE: send reinit to bring down nve1 - nve <<< NVE goes down %NVE-5-NVE_INTF_STATE: nve1: NVE Interface state changed to down From here, VXLAN traffic stops and all the devices behind the vPC will have traffic blocked for an extended duration Conditions: VPC VXLAN VTEP. SW1 reloads and is still running NVE source hold-time expiry. VPC peer-link is already up and both peers can see each other. Reload SW2 before NVE source hold-time expiry timer finishes. Traffic loss starts here while VPC leg is up on SW1 while NVE and source interface remains down. Loss persists till SW2 comes back and joins VPC and finishes all timers for VPC recovery and NVE interface.
Bug details contain sensitive information and therefore require a Cisco.com account to be viewed.
Bug Details Include
- Full Description (including symptoms, conditions and workarounds)
- Known Fixed Releases
- Related Community Discussions
- Number of Related Support Cases