Kubeless function invoke error

Question

Im having trouble invoking functions using kubeless. Here is the function spec

---
apiVersion: kubeless.io/v1beta1
kind: Function
metadata:
  name: smk
  namespace: smktest
spec:
  handler: hello.handler
  runtime: python2.7
  function: |
      import json
      def handler():
              return "hello world"
  deployment:
    spec:
      template:
        spec:
          containers:
             - env:
               - name: FOO
                 value: bar
               name: "smk-deployment"
               resources:
                 limits:
                   cpu: 100m
                   memory: 100Mi
                 requests:
                   cpu: 100m
                   memory: 100Mi

When I try to call the function as below,

kubeless function call smk 

I get

FATA[0000] Unable to find the service for smk

Two part question

1. How do I expose my function as a service
2. How do I specify Environment variables needed by this function ? Thank you

Update Running kubeless function ls --namespace=smktest yields below

NAME    NAMESPACE   HANDLER         RUNTIME     DEPENDENCIES    STATUS
smk     smktest     hello.handler   python2.7                   MISSING: Check controller logs

Next I tried kubectl logs -n kubeless -l kubeless=controller there's tons of error logs but I don't see anything specific to this function

Answer 1

When I try to call the function as below,

kubeless function call smk

I get

FATA[0000] Unable to find the service for smk

Running kubeless function ls --namespace=smktest

Then surely you would need to include the --namespace=smktest in your invocation command, too:

kubeless function call --namespace=smktest smk 

---

How do I specify Environment variables needed by this function ? Thank you

As best I can tell, there seems to be two approaches in use:

• Provide a Deployment template, which the function controller appears to merge but as far as I know container: image: is required in a Deployment, so you'd have to specify one in order to get access to its env: declaration
• Otherwise "cheat" and use the Pods ServiceAccount token to request cluster resources manually which might include a ConfigMap, Secret, or even resolving your own kubeless.io/function manifest and pulling something out of its annotations or similar