envoyproxy instance svc update error after gateway restart

[spwangxp]

Description:

What issue is being seen?
after gateway restart, gateway will update the exist envoy proxy instance, so there will encounter an error, seem like https://github.com/envoyproxy/gateway/blob/main/internal/infrastructure/kubernetes/infra.go#L74 miss add the original info to new svc.

Repro steps:

apply an gateway cr and envoyproxy cr, check an envoy proxy instance whether created in envoy-gateway-system, then restart the gateway.

Environment:

the main branch code;
k8s server: v1.19.8

Logs:

2023-04-23T11:00:59.473+0800 ERROR runner/runner.go:68 failed to create new infra {"runner": "infrastructure", "error": "failed to update service envoy-gateway-system/envoy-default-eg-64656661: Service "envoy-default-eg-64656661" is invalid: spec.clusterIP: Invalid value: "": field is immutable"}
github.com/envoyproxy/gateway/internal/infrastructure/runner.(*Runner).subscribeToProxyInfraIR.func1
gateway/internal/infrastructure/runner/runner.go:68
gateway/internal/message.HandleSubscription[...]
gateway/internal/message/watchutil.go:28
github.com/envoyproxy/gateway/internal/infrastructure/runner.(*Runner).subscribeToProxyInfraIR
gateway/internal/infrastructure/runner/runner.go:57

[spwangxp]

this if always be true due to svc.spec.ClusterIP/svc.spec.CheckNodePort/svc.spec.*.nodePort always different.

[zirain]

cc @arkodg , this seems a issue from the begining.