某局点S125G R7596P09三层聚合子接口BFD异常震荡问题

设备型号为S12504G-AF，版本为R7596P09设备，现场两台S125G堆叠，1框2框分别通过聚合对接华为设备，S125G与华为设备通过三层聚合子接口互联，建立BGP邻居，并配置BGP与BFD联动，相关配置如下：

bgp 125000

 #

 ip vpn-instance 5G_MEC

  peer 1.39.8.245 as-number 65398

  peer 1.39.8.245 description To-HNLT-RD-MCR-5G_MEC

  peer 1.39.8.245 bfd single-hop

  peer 1.39.8.249 as-number 65398

  peer 1.39.8.249 description To-HNLT-HX-MCR-5G_MEC

  peer 1.39.8.249 bfd single-hop

  #

  address-family ipv4 unicast

   aggregate 1.39.8.0 255.255.255.0

   import-route direct

   peer 1.39.8.245 enable

   peer 1.39.8.245 route-policy MCR_v4_5G_MEC_OUT export

   peer 1.39.8.249 enable

   peer 1.39.8.249 route-policy MCR_v4_5G_MEC_OUT export

现场发现BFD会话持续震荡：

%Apr 29 08:28:26:501 2022 S12500G_IRF2 BGP/5/BGP_STATE_CHANGED: BGP.MEC_N6: 100.64.0.37  state has changed from OPENCONFIRM to ESTABLISHED.

%Apr 29 08:28:26:506 2022 S12500G_IRF2 BFD/5/BFD_CHANGE_FSM: Sess[100.64.0.38/100.64.0.37, LD/RD:4339/34855, Interface:RAGG1.4094, SessType:Ctrl, LinkType:INET], Ver:1, Sta: DOWN->INIT, Diag: 0 (No Diagnostic)

%Apr 29 08:28:27:563 2022 S12500G_IRF2 BFD/5/BFD_CHANGE_FSM: Sess[100.64.0.38/100.64.0.37, LD/RD:4339/34855, Interface:RAGG1.4094, SessType:Ctrl, LinkType:INET], Ver:1, Sta: INIT->UP, Diag: 0 (No Diagnostic)

%Apr 29 08:28:28:061 2022 S12500G_IRF2 BGP/5/BGP_STATE_CHANGED: BGP.5G_MEC: 1.39.8.245  state has changed from OPENCONFIRM to ESTABLISHED.

%Apr 29 08:28:28:062 2022 S12500G_IRF2 BFD/5/BFD_CHANGE_FSM: Sess[1.39.8.246/1.39.8.245, LD/RD:4340/251662359, Interface:RAGG2.839, SessType:Ctrl, LinkType:INET], Ver:1, Sta: DOWN->UP, Diag: 0 (No Diagnostic)

%Apr 29 08:28:28:091 2022 S12500G_IRF2 BFD/5/BFD_CHANGE_FSM: Sess[1.39.8.250/1.39.8.249, LD/RD:4336/34853, Interface:RAGG1.839, SessType:Ctrl, LinkType:INET], Ver:1, Sta: UP->DOWN, Diag: 3 (Neighbor Signaled Session Down)

%Apr 29 08:28:28:100 2022 S12500G_IRF2 BGP/5/BGP_STATE_CHANGED: BGP.5G_MEC: 1.39.8.249  state has changed from ESTABLISHED to IDLE for session down event received from BFD.

%Apr 29 08:28:28:100 2022 S12500G_IRF2 BGP/5/BGP_STATE_CHANGED_REASON: BGP.5G_MEC: 1.39.8.249  state has changed from ESTABLISHED to IDLE. (Reason: a session down event was received from BFD, Error code: Send Notificationcode 6/0)

%Apr 29 08:28:28:163 2022 S12500G_IRF2 BFD/5/BFD_CHANGE_FSM: Sess[1.90.48.254/1.90.48.253, LD/RD:4334/34854, Interface:RAGG1.837, SessType:Ctrl, LinkType:INET], Ver:1, Sta: UP->DOWN, Diag: 3 (Neighbor Signaled Session Down)

原因是BFD存在配置限制：三层聚合接口/子接口不支持与BFD功能联动。

虽然通过修改设备模式为expert专家模式后，设备可以开启硬件BFD功能，从而支持与三层聚合子接口联动，现场修改设备工作模式并保存配置重启后发现，主框的BFD会话正常，不会震荡，但是备框的BFD会话则一直无法UP：

1框为主时：

<S12500G_IRF2>dis bfd session

 Total Session Num: 7     Up Session Num: 3     Init Mode: Active

 IPv4 session working in control packet mode:

 LD/RD          SourceAddr      DestAddr        State    Holdtime    Interface

 4097/0         11.11.11.1      11.11.11.2      Down        /        RAGG63

 4098/34817     100.64.0.38     100.64.0.37     Up       2000ms      RAGG1.4094

 4099/34818     1.39.8.250      1.39.8.249      Up       2000ms      RAGG1.839

 4100/34819     1.90.48.254     1.90.48.253     Up       2000ms      RAGG1.837

 4101/0         100.64.0.34     100.64.0.33     Down        /        RAGG2.4094

 4102/0         1.39.8.246      1.39.8.245      Down        /        RAGG2.839

 4103/0         1.90.48.250     1.90.48.249     Down        /        RAGG2.837

2框为主时：

<S12500G_IRF2>dis bfd session

 Total Session Num: 7     Up Session Num: 3     Init Mode: Active

 IPv4 session working in control packet mode:

 LD/RD          SourceAddr      DestAddr        State    Holdtime    Interface

 4104/0         11.11.11.2      11.11.11.1      Down        /        RAGG63

 4105/0         100.64.0.38     100.64.0.37     Down        /        RAGG1.4094

 4106/0         1.39.8.250      1.39.8.249      Down        /        RAGG1.839

 4107/0         1.90.48.254     1.90.48.253     Down        /        RAGG1.837

 4108/251662337 1.39.8.246      1.39.8.245      Up       2000ms      RAGG2.839

 4109/251662338 1.90.48.250     1.90.48.249     Up       2000ms      RAGG2.837

 4110/251662339 100.64.0.34     100.64.0.33     Up       2000ms      RAGG2.4094

修改为专家模式后，虽然能支持聚合子接口，但也存在限制，不支持VPN。硬件BFD下会选举一块接口板做为会话维护板，现网1框为主时选的是1框堆叠板，2框为主时选的是2框堆叠板，对于维护板自己的会话，底层有VPN信息，BFD状态正常，而对于非维护板，BFD报文需要送到维护板处理，由于上述限制会丢失VPN信息，导致BFD状态异常。

建议现网使用默认standard模式（注意调整系统模式需要保存配置并整框重启），即软件BFD实现方式，并通过vlan方式对接。