You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The frr implementation of metallb rejects any duplicate BGPPeer (i.e., peers that have the same IP for the same VRF).
Ideally, this uniqueness should be enforced on per node basis. Because it was impossible to detect (and maintain) upfront (and because it was the early stage of the frr implementation), we decided to block any overlapping IPs regardless of the fact that they might influence the behaviour of a distinct set of nodes.
This prevents some useful features, such as using different ASN / SRC address per nodes / set of nodes, or having custom timers per nodes (because a set of nodes could be on a different region with different latencies).
Describe the solution you'd like
What we propose here is to relax the constraint, trying to intercept misconfigurations in a best effort manner (with the webhooks) and blacklisting the neighbor ip (acting as if it was not configured at all) whenever there are multiple bgppeers eligible for the same node (on the same vrf), basically relying on the user to do the right configuration and resulting in a configuration error if this is not the case.
The user feedback should also improve whenever we implement #2156
This should also be extended to frr-k8s.
Additional context
No response
I've read and agree with the following
I've checked all open and closed issues and my request is not there.
I've checked all open and closed pull requests and my request is not there.
The text was updated successfully, but these errors were encountered:
is there any update on this? I would like to help in this regard but i think it would take me way longer to implement this than someone of the core team.
Is your feature request related to a problem?
The frr implementation of metallb rejects any duplicate BGPPeer (i.e., peers that have the same IP for the same VRF).
Ideally, this uniqueness should be enforced on per node basis. Because it was impossible to detect (and maintain) upfront (and because it was the early stage of the frr implementation), we decided to block any overlapping IPs regardless of the fact that they might influence the behaviour of a distinct set of nodes.
This prevents some useful features, such as using different ASN / SRC address per nodes / set of nodes, or having custom timers per nodes (because a set of nodes could be on a different region with different latencies).
Describe the solution you'd like
What we propose here is to relax the constraint, trying to intercept misconfigurations in a best effort manner (with the webhooks) and blacklisting the neighbor ip (acting as if it was not configured at all) whenever there are multiple bgppeers eligible for the same node (on the same vrf), basically relying on the user to do the right configuration and resulting in a configuration error if this is not the case.
The user feedback should also improve whenever we implement #2156
This should also be extended to frr-k8s.
Additional context
No response
I've read and agree with the following
The text was updated successfully, but these errors were encountered: