tnozicka/openshift-acme

Hardcoded limits (openshift-acme-exposer)

rbo opened this issue · 2 comments

rbo commented

What would you like to be added:

Please make ResourceRequirements for exposer configurable: https://github.com/tnozicka/openshift-acme/blob/master/pkg/controller/route/route.go#L868

via --exposer-resource-cpu... or via Env variable.

Why is this needed:
On OpenShift Online if have a limit of 80Mi :

 oc describe limits/resource-limits
Name:                  resource-limits
Namespace:             sandbox
Type                   Resource  Min   Max  Default Request  Default Limit  Max Limit/Request Ratio
----                   --------  ---   ---  ---------------  -------------  -----------------------
Pod                    cpu       5m    6    -                -              -
Pod                    memory    80Mi  3Gi  -                -              -
Container              cpu       5m    6    30m              1              -
Container              memory    80Mi  3Gi  409Mi            512Mi          -
PersistentVolumeClaim  storage   1Gi   -    -                -              -

And I can not change the limit because it is a shared cluster.

That means the pods from the e openshift-acme-exposer will never start because if:

 invalid: spec.containers[0].resources.requests: Invalid value: "80Mi": must be less than or equal to memory limit
  Warning  FailedCreate  2m54s (x9 over 9m2s)  replicaset-controller  (combined from similar events): Error creating: Pod "exposer-hmohfkv2ungr7g46qfchpr2e1if
n9fk1jttt846sqdd0prl65f4pk4h" is invalid: spec.containers[0].resources.requests: Invalid value: "80Mi": must be less than or equal to memory limit

@tnozicka

I suppose we could possibly list the limits and adjust the min value automatically without exposing a flag.

rbo commented

If this is easier, up to you...