For enabling TLS, as new application port is involved the service should be redeployed (by modifying the parameters in blueprint), esp when changing from noAuth to basicAuth/certOnly/certBasicAuth. The latter 3 types use 8443 while the noAuth uses 8080. Any changes within basicAuth/certOnly/certBasicAuth can be done through consul update as k8s deployment descriptor (which contains the service definition and healthcheck spec) are still valid.
High level test cases for auth.method = "basicAuth" :
When application is setup for TLS and auth.method = "basicAuth", healthcheck endpoint must be using 8443 (change submitted to override and support 8080 for healthcheck is not required nor valid).
To change VES Collector flag to basic.auth and adopt healthhecks to use HTTPS, there is need to change VES blueprints. Steps:
- Copy attached k8s-ves-0529-secure.yaml file to ONAP Bootstrap node
- From Rancher node find bootstrap pod: kubectl -n onap get pod | grep -i bootstrap
- Login to found in point 1 pod: kubectl exec -it <bootstrap_pod_name> -n onap bash
- Uninstal VES: cfy uninstall ves
- Delete VES blueprint: cfy blueprints delete ves
- Overwrite VES default blueprint with copied k8s-ves-0529-secure.yaml: copy new bp - k8s-ves-0529-secure.yaml as /blueprints/k8s-ves.yaml
- Apply new blueprint: cfy install -b ves -d ves -i /inputs/k8s-ves-inputs.yaml /blueprints/k8s-ves.yaml
TC ID | Test Case Name | Test Case Description | Expected Result |
---|---|---|---|
T01 | Client with correct basic auth and correct certificate | PASS | |
T02 | Client with correct basic auth and incorrect certificate | PASS | |
T03 | Client with correct basic auth and without certificate | PASS | |
T04 | Client without basic auth and without certificate | FAIL | |
T05 | Client without basic auth and with correct certificate | FAIL | |
T06 | Client with incorrect basic auth and with correct certificate | FAIL | |
T07 | Client with incorrect basic auth and with correct certificate | FAIL | |