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
Is your feature request related to a problem? Please describe.
Consider IPv6 only cluster where Attractor relies on network-attachment-definition to provide interface for cluster breakout:
The NAD takes care of IP address allocation through some IPAM of choice to assign IPv6 address only. Howver, due to lack of any feasible IPv4 addresses, BIRD in LB-FE will bail out with following error: "Cannot determine router ID, please configure it manually"
Thus, FE containers get stuck in CrashLoopBackOff.
That's because the configuration generated for BIRD does not contain a router ID. Having unique router IDs would require some IPAM like feature that would assign router IDs to FE instances.
The difference compared to VLAN-NSE configuration is that in case of NSM provided external interface, IPv4 addresses were always assigned by default.
As a workaround, the IPAM config part in the NAD could still assign an IPv4 address that could be used by BIRD as router ID.
Describe the solution you'd like
Consider if router ID should be provided independently from the external interface's configuration.
Describe alternatives you've considered
Some time ago a tested with a router ID shared among the FEs of the same Attractor. The peers routers also running BIRD seemed to have no problem with that. Although I haven't done extensive testing.
Additional context
N/A
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Consider IPv6 only cluster where Attractor relies on network-attachment-definition to provide interface for cluster breakout:
The NAD takes care of IP address allocation through some IPAM of choice to assign IPv6 address only. Howver, due to lack of any feasible IPv4 addresses, BIRD in LB-FE will bail out with following error: "Cannot determine router ID, please configure it manually"
Thus, FE containers get stuck in CrashLoopBackOff.
That's because the configuration generated for BIRD does not contain a router ID. Having unique router IDs would require some IPAM like feature that would assign router IDs to FE instances.
The difference compared to VLAN-NSE configuration is that in case of NSM provided external interface, IPv4 addresses were always assigned by default.
As a workaround, the IPAM config part in the NAD could still assign an IPv4 address that could be used by BIRD as router ID.
Describe the solution you'd like
Consider if router ID should be provided independently from the external interface's configuration.
Describe alternatives you've considered
Some time ago a tested with a router ID shared among the FEs of the same Attractor. The peers routers also running BIRD seemed to have no problem with that. Although I haven't done extensive testing.
Additional context
N/A
The text was updated successfully, but these errors were encountered: