Ambassador: Ambassador 0.70.0 Pod Panics

[dpaks]

Describe the bug
After updating to 0.70.0, my ambassador pod crashed.

To Reproduce
Steps to reproduce the behavior:

1. Update to 0.70.0
2. Ambassador pod will go into crashloopbackoff

Expected behavior
A running ambassador pod

Versions (please complete the following information):

• Ambassador: 0.70.0
• Kubernetes environment: K8s cluster on AWS, Ubuntu 18.04
• Version: kubernetes 1.12.5

Additional context

2019/05/20 11:58:16 consulwatchman: signaling shutdown
2019/05/20 11:58:16 kubewatchman: signaling shutdown
2019/05/20 11:58:16 aggregator: signaling shutdown
2019/05/20 11:58:16 invoker: signaling shutdown
2019/05/20 11:58:16 api: signaling shutdown
2019/05/20 11:58:16 api[1]: signaling shutdown
2019/05/20 11:58:16 consulwatchman: shutdown initiated
2019/05/20 11:58:16 kubewatchman: shutdown initiated
2019/05/20 11:58:16 invoker: shutdown initiated
2019/05/20 11:58:16 api: listener close error: close tcp [::]:8002: use of closed network connection
2019/05/20 11:58:16 api[1]: http: Server closed
2019/05/20 11:58:16 kubebootstrap: WORKER PANICKED: unrecognized resource: AuthService
goroutine 155 [running]:
runtime/debug.Stack(0xc0005258b8, 0x118e3a0, 0xc0001f2450)
/usr/local/go/src/runtime/debug/stack.go:24 +0xa7

github.com/datawire/teleproxy/pkg/supervisor.(Supervisor).launch.func1.1.1(0xc000525f88)
/home/circleci/repo/pkg/supervisor/supervisor.go:305 +0x60
panic(0x118e3a0, 0xc0001f2450)
/usr/local/go/src/runtime/panic.go:513 +0x1b9
github.com/datawire/teleproxy/pkg/k8s.(Client).ResolveResourceType(0xc00000cc60, 0xc0004043b0, 0xb, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, ...)
/home/circleci/repo/pkg/k8s/client.go:232 +0x6c5
github.com/datawire/teleproxy/pkg/k8s.(Watcher).SelectiveWatch(0xc0006a6c40, 0x0, 0x0, 0xc0004043b0, 0xb, 0x0, 0x0, 0x0, 0x0, 0xc0002721c0, ...)
/home/circleci/repo/pkg/k8s/watcher.go:134 +0x60
main.(kubebootstrap).Work(0xc000646930, 0xc00000cd00, 0xc00006bf01, 0xc00006bf28)
/home/circleci/repo/cmd/watt/kubewatchman.go:138 +0x325
main.(kubebootstrap).Work-fm(0xc00000cd00, 0x13f5630, 0xc00006bf88)
/home/circleci/repo/cmd/watt/main.go:107 +0x34
github.com/datawire/teleproxy/pkg/supervisor.(Supervisor).launch.func1.1(0xc000525f88, 0xc000646a10, 0xc00000cd00)
/home/circleci/repo/pkg/supervisor/supervisor.go:310 +0x70
github.com/datawire/teleproxy/pkg/supervisor.(Supervisor).launch.func1(0xc000646a10, 0xc00000cd00, 0xc0006469a0)
/home/circleci/repo/pkg/supervisor/supervisor.go:311 +0x51
created by github.com/datawire/teleproxy/pkg/supervisor.(Supervisor).launch
/home/circleci/repo/pkg/supervisor/supervisor.go:300 +0xb2

2019/05/20 11:58:16 api[1]: http: Server closed
AMBASSADOR: watt exited with status 1
AMBASSADOR: shutting down (1)

[richarddli]

Did you upgrade to 0.70 following the instructions in the blog post?

https://blog.getambassador.io/announcing-ambassador-0-70-app-modernization-beyond-kubernetes-9d366062b395

[volatilemolotov]

I have upgraded using helm upgrade and got the same error

[dpaks]

Did you upgrade to 0.70 following the instructions in the blog post?

https://blog.getambassador.io/announcing-ambassador-0-70-app-modernization-beyond-kubernetes-9d366062b395

I followed and it worked. Just had to apply https://www.getambassador.io/yaml/ambassador/ambassador-rbac.yaml with custom namespace change.

[davidharkis]

I'm getting this issue while trying to use 0.70.0 without RBAC on Openshift 3.9. We do not have cluster level control so cannot create CRDs. I'm struggling to see any other differences in the YAML that would cause this.

[przemek-sl]

Is it even possible to deploy 0.70 without RBAC in k8s?

[nbkrause]

@davidharkis The CRDs are required in 0.70.0. If there is no way for you to create them. We will need to release a patch so they are not required for Ambassador to start.

@przemek-sl It is possible to deploy without RBAC but the CRDs are required

[kflynn]

0.70.1 should take care of this.