Idcmp

Caught one: ``` # kubectl logs -f -n kube-system kilo-7mkm8 Defaulted container "kilo" out of: kilo, install-cni (init) {"caller":"main.go:280","msg":"Starting Kilo network mesh '0.5.0'.","ts":"2022-05-05T18:09:35.628352703Z"} {"caller":"cni.go:61","component":"kilo","err":"failed to read IPAM config from CNI...

Okay, I'm still experiencing wireguard dropouts, (as described in Situation), but both nodes think they're `Ready` now. I restarted the laptop VM so possibly there was some junk network configuration...

> It's quite funny that Kilo is getting terminated. It seems that Kubernetes is shutting it down for some reason. Did you by chance add liveness checks to the daemonset?...

Added --log-level=debug to the kilo daemonset, will post an update when it dies.

Ran `chronyc` first so both machines are within a second of each other: Crashing kilo: ``` {"caller":"mesh.go:280","component":"kilo","event":"update","level":"debug","msg":"syncing nodes","ts":"2022-05-05T21:37:56.104998312Z"} {"DiscoveredEndpoints":{},"caller":"mesh.go:817","component":"kilo","level":"debug","msg":"Discovered WireGuard NAT Endpoints","ts":"2022-05-05T21:38:13.127679088Z"} {"allowedIPs":[{"IP":"10.244.0.0","Mask":"////AA=="},{"IP":"ec2.private.ip","Mask":"/////w=="}],"allowedLocationIPs":null,"caller":"topology.go:191","cidrs":[{"IP":"10.244.0.0","Mask":"////AA=="}],"component":"kilo","endpoint":"ec2.public.ip:51820","hostnames":["ip-.compute.internal"],"leader":0,"level":"debug","location":"location:aws","msg":"generated segment","privateIPs":["ec2.private.ip"],"ts":"2022-05-05T21:38:13.127713095Z"} {"allowedIPs":[{"IP":"10.244.1.0","Mask":"////AA=="},{"IP":"laptop.vm.private.ip","Mask":"/////w=="}],"allowedLocationIPs":null,"caller":"topology.go:191","cidrs":[{"IP":"10.244.1.0","Mask":"////AA=="}],"component":"kilo","endpoint":"laptop.vm.private.ip:51820","hostnames":["fedora"],"leader":0,"level":"debug","location":"location:laptop","msg":"generated segment","privateIPs":["laptop.vm.private.ip"],"ts":"2022-05-05T21:38:13.127785181Z"} {"caller":"topology.go:239","component":"kilo","hostname":"fedora","leader":true,"level":"debug","location":"location:laptop","msg":"generated topology","privateIP":"laptop.vm.private.ip/24","subnet":"10.244.1.0/24","ts":"2022-05-05T21:38:13.127804122Z","wireGuardIP":"10.4.0.2/16"} {"caller":"mesh.go:368","component":"kilo","level":"debug","msg":"successfully checked...

@squat is there anything i can do to help?

Just for tracking purposes, this _doesn't_ fix the client-half regression mentioned in https://github.com/golang/go/issues/56342

Is it just money, or is Apple against something like WavTap?

I don't mind providing more information, more detailed logs or testing patches. I'm just unsure if this project is maintained anymore.