Fix e2e test when pod runs in openshift with correct security configuration

[claudio4j]

The e2e test sets the namespace to a restricted PSA.
When the integration is run in openshift, it sets the correct securityContext to run with no privileges:

securityContext:
  allowPrivilegeEscalation: false
  capabilities:
    drop:
    - ALL
  runAsNonRoot: true
  runAsUser: 1000650000
  seccompProfile:
    type: RuntimeDefault

The pod runs just fine with no errors.
So, this is just to disable this test in openshift.

Release Note

NONE

[squakez]

This test was done on purpose to fix a wrong behavior in Openshift, see #4974 - it would be good to understand if/how to rewrite it in order to simulate the failing conditions in OC instead of skipping.

[claudio4j]

Thanks for pointing the original issue, I will further investigate it.

[claudio4j]

I could reproduce the error in #4974 by installing camel-k 2.1.0 from operator hub in openshift 4.15.14.
Updated to Camel K 2.2.0 from operator hub.
Then there is no error anymore.
The fix was provided by #4932
The problem was in 2.1.0, the runAsUser field of securityContext was set from the platform namespace, while the correct one should be from the integration namespace.

The proposal of the test is to run an integration in a restricted namespace, whereas the restricted policy is the more secured one by the policy definition.
However when running in openshift the operator correctly sets the securityContext in compliance with the restricted policy, whereas when running not on openshift profile, the securityContext is not applied, so that results in error when the namespace is set with the restricted policy.
I will open an issue to set the securityContext policy regardless if it's running in openshift, this way the integration can correctly run in a restricted policy in any kubernetes cluster.

[squakez]

Thanks for looking at it. This is something we'll be already enforcing from version 2.4 onward: #5461