Output from make release fails with cabpt-controller-manager going into crashloop

ag3man commented 3 years ago

cloned repo switched to v0.2.0-alpha-11

make release
kubectl apply -f _out/bootstrap-components.yaml

(⎈ |personal-cluster:default)➜  cluster-api-bootstrap-provider-talos git:(9c5538c) kg pod -A
NAMESPACE                           NAME                                                             READY   STATUS             RESTARTS   AGE
cabpt-system                        cabpt-controller-manager-57fcd6486f-hqzwh                        0/1     CrashLoopBackOff   1          9s
capa-system                         capa-controller-manager-c954d99bb-nt7t4                          2/2     Running            0          6h17m
capi-kubeadm-bootstrap-system       capi-kubeadm-bootstrap-controller-manager-6c6fd4977-r8w2f        2/2     Running            0          6h17m
capi-kubeadm-control-plane-system   capi-kubeadm-control-plane-controller-manager-d6598bf4-9ttlp     2/2     Running            0          6h17m
capi-system                         capi-controller-manager-6c78fff988-ptxlf                         2/2     Running            0          6h17m
capi-webhook-system                 cabpt-controller-manager-75bbc79fc9-rjn29                        1/1     Running            0          9s
capi-webhook-system                 capa-controller-manager-6c488c5d48-56v98                         2/2     Running            0          6h17m
capi-webhook-system                 capi-controller-manager-868796f585-txml9                         2/2     Running            0          6h17m
capi-webhook-system                 capi-kubeadm-bootstrap-controller-manager-765c848f8-vpf2m        2/2     Running            0          6h17m
capi-webhook-system                 capi-kubeadm-control-plane-controller-manager-6769d967bc-t8zmj   2/2     Running            0          6h17m
cert-manager                        cert-manager-86cb5dcfdd-fd2fx                                    1/1     Running            0          6h18m
cert-manager                        cert-manager-cainjector-84cf775b89-mskn4                         1/1     Running            0          6h18m
cert-manager                        cert-manager-webhook-5d5dc765f6-b4ppw                            1/1     Running            0          6h18m
kube-system                         cilium-9c22r                                                     1/1     Running            0          6h26m
kube-system                         cilium-bfb5l                                                     1/1     Running            0          6h26m
kube-system                         cilium-operator-7c755f4594-2cmb8                                 1/1     Running            0          6h26m
kube-system                         coredns-74ff55c5b-6swdr                                          1/1     Running            0          6h26m
kube-system                         etcd-personal-cluster                                            1/1     Running            0          6h26m
kube-system                         kube-apiserver-personal-cluster                                  1/1     Running            0          6h26m
kube-system                         kube-controller-manager-personal-cluster                         1/1     Running            0          6h26m
kube-system                         kube-proxy-2vhjb                                                 1/1     Running            0          6h26m
kube-system                         kube-proxy-cxxnl                                                 1/1     Running            0          6h26m
kube-system                         kube-scheduler-personal-cluster                                  1/1     Running            0          6h26m
kube-system                         storage-provisioner                                              1/1     Running            0          6h26m


(⎈ |personal-cluster:default)➜  cluster-api-bootstrap-provider-talos git:(9c5538c) kl -n cabpt-system pods/cabpt-controller-manager-57fcd6486f-hqzwh
invalid argument "MachinePool=${EXP_MACHINE_POOL:=false}" for "--feature-gates" flag: invalid value of MachinePool=${EXP_MACHINE_POOL:=false}, err: strconv.ParseBool: parsing "${EXP_MACHINE_POOL:=false}": invalid syntax
invalid argument "MachinePool=${EXP_MACHINE_POOL:=false}" for "--feature-gates" flag: invalid value of MachinePool=${EXP_MACHINE_POOL:=false}, err: strconv.ParseBool: parsing "${EXP_MACHINE_POOL:=false}": invalid syntax
Usage of /manager:
      --enable-leader-election        Enable leader election for controller manager. Enabling this will ensure there is only one active controller manager.
      --feature-gates mapStringBool   A set of key=value pairs that describe feature gates for alpha/experimental features. Options are:
                                      AllAlpha=true|false (ALPHA - default=false)
                                      AllBeta=true|false (BETA - default=false)
                                      ClusterResourceSet=true|false (ALPHA - default=false)
                                      MachinePool=true|false (ALPHA - default=false)
      --kubeconfig string             Paths to a kubeconfig. Only required if out-of-cluster.
      --master --kubeconfig           (Deprecated: switch to --kubeconfig) The address of the Kubernetes API server. Overrides any value in kubeconfig. Only required if out-of-cluster.
      --metrics-addr string           The address the metric endpoint binds to. (default ":8080")
      --webhook-port int              Webhook Server port, disabled by default. When enabled, the manager will only work as webhook server, no reconcilers are installed.

rsmitty commented 3 years ago

Hey @ag3man, not quite a bug directly, but it's expected that release files will be applied from the repo with clusterctl. So if you're rolling your own release, you've probably got a few options:

Just remove the MachinePool feature gate, or set it to your desired value manually.
Pipe the file through envsubst before applying. Something like cat _out/bootstrap-components.yaml | envsubst | kubectl apply -f -
Add some kustomize that wraps our stuff to do ^ for you automatically.
Take a look at the upstream clusterctl docs and configure a "local" repository for clusterctl to use and do a clusterctl init -b talos. This will ensure that env vars are substituted.

After typing all that, the 2nd one is probably the one that makes the most sense imo. Hope this helps!

ag3man commented 3 years ago

Thanks for the response @rsmitty I went the manual way and edited manifest to override and set MachinePool=true.

I will close this as invalid.

siderolabs / cluster-api-bootstrap-provider-talos

Output from make release fails with cabpt-controller-manager going into crashloop #59