Seldon Manager Config Map has an extra comma that can cause parsing error

[gangpeng]

Describe the bug

Seldon Manager config map has an extra "," in prediction servers section at following line:

seldon-core/operator/config/manager/configmap.yaml

Line 83 in b29686a

},

This causes the webhook to panic with following stack:
/usr/local/go/src/net/http/server.go:2993 +0x39b
{"level":"info","ts":1633731619.884156,"logger":"seldondeployment","msg":"Validating v1 Webhook called for CREATE","name":"iris-model"}
2021/10/08 22:20:19 http: panic serving 172.17.0.1:1969: Unable to unmarshall predictor_servers
json string due to invalid character '}'
looking for beginning of object key string
goroutine 6050 [running]:
net/http.(*conn).serve.func1(0xc000626640)
/usr/local/go/src/net/http/server.go:1804 +0x153
panic(0x1a13620, 0xc000a01160)
/usr/local/go/src/runtime/panic.go:971 +0x499

github.com/seldonio/seldon-core/operator/apis/machinelearning.seldon.io/v1.getPredictorServerConfigsFromMap(0xc00043af00, 0x1ec6f88, 0xc000120008, 0xc00004616e)
/workspace/apis/machinelearning.seldon.io/v1/prepack.go:143 +0x1d7
github.com/seldonio/seldon-core/operator/apis/machinelearning.seldon.io/v1.getPredictorServerConfigs(0x203000, 0x7fc1ecda2ce0, 0x3f)
/workspace/apis/machinelearning.seldon.io/v1/prepack.go:135 +0x249
github.com/seldonio/seldon-core/operator/apis/machinelearning.seldon.io/v1.(*SeldonDeploymentSpec).checkPredictiveUnits(0xc0005b67d8, 0xc0005164c0, 0xc0005164b0, 0xc0007a8bd0, 0x0, 0x0, 0x0, 0xc000516208, 0x4dbf53, 0xc0001340c0)
/workspace/apis/machinelearning.seldon.io/v1/seldondeployment_webhook.go:100 +0xc8b
github.com/seldonio/seldon-core/operator/apis/machinelearning.seldon.io/v1.(*SeldonDeploymentSpec).ValidateSeldonDeployment(0xc0005b67d8, 0x1ca3d90, 0x27)
/workspace/apis/machinelearning.seldon.io/v1/seldondeployment_webhook.go:251 +0x573
github.com/seldonio/seldon-core/operator/apis/machinelearning.seldon.io/v1.(*SeldonDeployment).ValidateCreate(0xc0005b66c0, 0xc00095bb00, 0x24)
/workspace/apis/machinelearning.seldon.io/v1/seldondeployment_webhook.go:290 +0xe5
sigs.k8s.io/controller-runtime/pkg/webhook/admission.(*validatingHandler).Handle(0xc0005974b8, 0x1ec6f50, 0xc000a96800, 0xc00095bb00, 0x24, 0xc0009a3e20, 0x19, 0xc000a15178, 0x2, 0xc000a15180, ...)
/go/pkg/mod/sigs.k8s.io/controller-runtime@v0.9.6/pkg/webhook/admission/validator.go:71 +0xe37
sigs.k8s.io/controller-runtime/pkg/webhook/admission.(*Webhook).Handle(0xc000e8c600, 0x1ec6f50, 0xc000a96800, 0xc00095bb00, 0x24, 0xc0009a3e20, 0x19, 0xc000a15178, 0x2, 0xc000a15180, ...)
/go/pkg/mod/sigs.k8s.io/controller-runtime@v0.9.6/pkg/webhook/admission/webhook.go:146 +0xd5
sigs.k8s.io/controller-runtime/pkg/webhook/admission.(*Webhook).ServeHTTP(0xc000e8c600, 0x7fc1ecd69d90, 0xc000930d20, 0xc00097cd00)
/go/pkg/mod/sigs.k8s.io/controller-runtime@v0.9.6/pkg/webhook/admission/http.go:99 +0x1045

To reproduce

1. Install Seldon on Minikube:
   $helm install seldon-core seldon-core-operator \
   --repo https://storage.googleapis.com/seldon-charts
   --set ambassador.enabled=false
   --set managerCreateResources=true
   --namespace seldon-system

2. Deploy Iris following Seldon doc:
   $kubectl create namespace seldon
   $ kubectl apply -f - << END
   apiVersion: machinelearning.seldon.io/v1
   kind: SeldonDeployment
   metadata:
   name: iris-model
   namespace: seldon
   spec:
   name: iris
   predictors:

• graph:
  implementation: SKLEARN_SERVER
  modelUri: gs://seldon-models/sklearn/iris
  name: classifier
  name: default
  replicas: 1
  END

I also created a go program to reproduce the error. Please open attached zip file and extract both files to the same directory and run "go predictor.go".

Expected behaviour

Controller should accept the deployment.

Environment

minikube version
minikube version: v1.22.0
commit: a03fbcf166e6f74ef224d4a63be4277d017bb62e
Archive.zip

[ukclivecox]

Thanks for this. Will look into it for next release