-
Bug
-
Resolution: Done
-
Blocker
-
None
-
False
-
False
-
Undefined
-
What problem/issue/behavior are you having trouble with? What do you expect to see?
When the operator creates the deployment for the postgres database used by clair, it uses an image that is different to the one configured following the official documentation [0].
I was trying to specify the postgres image I want the operator to use and discovered that there is already an image configured that will be okay for me to use in our environment (coming from
).
-> oc get csv quay-operator.v3.5.2 -o yaml | grep -A 1 RELATED_IMAGE_COMPONENT_POSTGRES
- name: RELATED_IMAGE_COMPONENT_POSTGRES
value:
The environment variable specified in the CSV is also used when deploying the operator pod:
-> oc get pod quay-operator.v3.5.2-6b6c7545f4-2bhxf -o yaml | grep -A 1 RELATED_IMAGE_COMPONENT_POSTGRES
- name: RELATED_IMAGE_COMPONENT_POSTGRES
value:
But when the operator creates the clair postgres deployment, it does not use this image and instead specifies the following:
"centos/postgresql-10-centos7@sha256:de1560cb35e5ec643e7b3a772ebaac8e3a7a2a8e8271d9e91ff023539b4dfb33"
which I am not able to use in my environment, since I only allow images provided by RedHat.
This operator is a RedHat product, so I also would assume that it will use RedHat images by default.
I've set the postgres image from
manually in the clair postgres deployment but it is reverted every time the operator updates or restarts, so this is not really a state im happy with.
[0]
Where are you experiencing the behavior? What environment?
OpenShift 4.6
Â
Acceptance criteria:
- Image used for Clair's Postgres deployment respects value of RELATED_IMAGE_COMPONENT_POSTGRES variable
- All postgres deployments created by the Operator uses rhel8 images