[FROG] About eliminating duplicate BFD peer session
吴昌容
wcr819 at gmail.com
Sat Jan 8 01:16:48 UTC 2022
Dear FRR community members,
I am a user of FRR and I am now using it to build a small-scale experiment
network. I ran into a problem regarding the interaction between BFD and BGP
in the recent days. My BGP configuration specifies BFD dependencies for BGP
neighbors using "neighbor X.X.X.X bfd". And the FRR will create some BFD
peer sessions (marked as dynamic in FRR) immediately when it reads the BGP
configuration. Then after FRR reads the BFD configuration, it seems that
FRR will deregister the previous "dynamic" peer sessions and register those
peer sessions (marked as configured) specified in the BFD configuration.
I learn this from the debugging log and such mechanisms work in most cases.
But in some situations, it does not work. Specifically, the FRR will have
two duplicate BFD peer sessions for the same peer at last. I found that
restarting FRR can solve the problem. Yet, I wonder whether there is a more
elegant solution.
The example configuration of the problematic peer is as follows:
router #1, with IP address 10.3.0.1
*router bgp 1 no bgp ebgp-requires-policy neighbor 10.0.0.2 remote-as
2 neighbor 10.0.0.2 bfd neighbor 10.1.0.2 remote-as 3 neighbor 10.1.0.2
bfd neighbor 10.4.0.1 remote-as 1 neighbor 10.4.0.1 bfd neighbor 10.4.0.1
update-source 10.3.0.1 neighbor 10.5.0.1 remote-as 1 neighbor 10.5.0.1
bfd neighbor 10.5.0.1 update-source 10.3.0.1 ! address-family ipv4 unicast
network 10.0.0.0/24 <http://10.0.0.0/24> network 10.1.0.0/24
<http://10.1.0.0/24> network 10.3.0.0/24 <http://10.3.0.0/24> network
10.4.0.0/24 <http://10.4.0.0/24> network 10.5.0.0/24 <http://10.5.0.0/24>
network 10.12.0.0/24 <http://10.12.0.0/24> network 10.13.0.0/24
<http://10.13.0.0/24> network 10.18.0.0/24
<http://10.18.0.0/24> exit-address-familyexitbfd peer 10.0.0.2
transmit-interval 100 receive-interval 100 exit ! peer 10.1.0.2
transmit-interval 100 receive-interval 100 exit ! peer 10.4.0.1 multihop
local-address 10.3.0.1 transmit-interval 100 receive-interval
100 exit ! peer 10.5.0.1 multihop local-address 10.3.0.1 transmit-interval
100 receive-interval 100 exit !exit*
router #2, with IP address 10.5.0.1
*router bgp 1 no bgp ebgp-requires-policy neighbor 10.3.0.1 remote-as
1 neighbor 10.3.0.1 bfd neighbor 10.3.0.1 update-source
10.5.0.1exitbfd peer 10.3.0.1 multihop local-address 10.5.0.1
transmit-interval 100 receive-interval 100 exit !exit*
The router #1 get two duplicate BFD peer sessions as follows: (the
interface s0-eth2 has the IP address 10.3.0.1)
*peer 10.5.0.1 local-address 10.3.0.1 vrf default interface s0-eth2
ID: 2040029068 Remote ID: 0 Active
mode Status: down Downtime: 11 minute(s), 51
second(s) Diagnostics: ok Remote diagnostics:
ok Peer Type: dynamic Local timers:
Detect-multiplier: 3 Receive interval:
300ms Transmission interval: 300ms
Echo receive interval: 50ms Echo transmission
interval: disabled Remote timers:
Detect-multiplier: 3 Receive interval: 1000ms
Transmission interval: 1000ms Echo
receive interval: disabled*
* peer 10.5.0.1 multihop local-address 10.3.0.1 vrf default
ID: 1516906735 Remote ID: 2466577291 Active
mode Minimum TTL: 254 Status: up
Uptime: 11 minute(s), 51 second(s) Diagnostics: ok
Remote diagnostics: ok Peer Type: configured
Local timers: Detect-multiplier: 3
Receive interval: 100ms Transmission
interval: 100ms Echo receive interval: 50ms
Echo transmission interval: disabled Remote
timers: Detect-multiplier: 3
Receive interval: 100ms Transmission interval:
100ms Echo receive interval: 50ms*
Can anyone give me a hint of a clean solution?
Thanks!
Changrong Wu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.frrouting.org/pipermail/frog/attachments/20220107/648162be/attachment.htm>
More information about the frog
mailing list