Details
-
Bug
-
Resolution: Won't Do
-
Major
-
None
-
2.12.1 GA
-
5
-
False
-
None
-
False
-
Not Started
-
Not Started
-
Not Started
-
Not Started
-
Not Started
-
Not Started
Description
Currently our documentation states 2 things to check:
In step 6:
oc get pods
in step 8:
oc get apimanager <myapimanager> -o yaml
Step 6 is not always true. Some pods might not roll out after an upgrade because it has the same image and configuration as the previous version. It would be good to mention in each release what is expected.
Step 8. It is not sufficient checking the annotations of the APImanager. Because in on case we saw that system-mysql deployment had been updated but the new version had failed to roll out. The old deployment is running instead with the previous version of the image. A customer can only check this by monitoring the pods and logs, but not from the APImanager. Therefore there needs to be a better set of steps that customers can rely on to verify that all deployments and the whole system has been upgraded as expected, if that means building some of those checks into the operator itself then that might be the better option but right now what is documented is insufficient.
The request is to have a better documented verification steps to check whether an upgrade is successful or not.