BGP restart not working with BFD

[pantxisto]

Hello, I have a kubernetes cluster with one master node and 2 worker nodes. This cluster uses metallb which uses FRR with BGP and BFD for loadbalancing. Per node it creates some speakers with the next BGP and BFD promerties and values:

peers:
    - peer-address: 10.0.0.1
      peer-asn: 64600
      my-asn: 64600
      bfd-profile: full1
    bfd-profiles:
    - name: full1
      receive-interval: 179
      transmit-interval: 180
      echo-interval: 62
      echo-mode: false
      passive-mode: false
      minimum-ttl: 254
    address-pools:
    - name: default
      protocol: bgp
      addresses:
      - 128.28.28.1-128.28.28.20
     

On the router I have the next configuration:

bfd
  profile full1
    receive-interval 179
    transmit-interval 150
    echo receive-interval 62
    echo transmit-interval 62
    minimum-ttl 254
  peer 10.0.0.2
    profile full1
    no shutdown
  peer 10.0.0.3
    profile full1
    no shutdown
  peer 10.0.0.4
    profile full1
    no shutdown
    
router bgp 64600
  bgp router-id 10.0.0.1
  neighbor 10.0.0.2 remote-as 64600
  neighbor 10.0.0.2 bfd
  neighbor 10.0.0.3 remote-as 64600
  neighbor 10.0.0.3 bfd
  neighbor 10.0.0.4 remote-as 64600
  neighbor 10.0.0.4 bfd

BGP and BFD connection is happening over a gre tunnel with 255 ttl.

When stopping worker nodes and with them speakers I get the next output on router when executing "show bfd peers" on vtysh:

BFD Peers:
	peer 10.0.0.4 vrf default
		ID: 2422357222
		Remote ID: 0
		Active mode
		Status: down
		Downtime: 11 minute(s), 35 second(s)
		Diagnostics: neighbor signaled session down
		Remote diagnostics: neighbor signaled session down
		Peer Type: configured
		Local timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 150ms
			Echo receive interval: 62ms
			Echo transmission interval: disabled
		Remote timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 180ms
			Echo receive interval: 62ms

	peer 10.0.0.3 vrf default
		ID: 144310512
		Remote ID: 0
		Active mode
		Status: down
		Downtime: 24 minute(s), 31 second(s)
		Diagnostics: control detection time expired
		Remote diagnostics: ok
		Peer Type: configured
		Local timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 150ms
			Echo receive interval: 62ms
			Echo transmission interval: disabled
		Remote timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 180ms
			Echo receive interval: 62ms

	peer 10.0.0.2 vrf default
		ID: 2275962565
		Remote ID: 545881773
		Active mode
		Status: up
		Uptime: 38 minute(s), 35 second(s)
		Diagnostics: ok
		Remote diagnostics: ok
		Peer Type: configured
		Local timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 150ms
			Echo receive interval: 62ms
			Echo transmission interval: disabled
		Remote timers:
			Detect-multiplier: 3
			Receive interval: 179ms
			Transmission interval: 180ms
			Echo receive interval: 62ms
			

After restarting both nodes and waiting 5 minutes the diagnostics of BFD peers says:

• worker1: control detection time expired.
• worker2: neighbor signaled session down.

I don't understand this messages. It is not supposed that BFD should communicate BGP which nodes are up or down? And when nodes are restarted should not be noticed to BFD?

Thanks.