Description of problem: If an image is already existed in the openshift pulled from a remote repository like docker,io, $ oc get is NAME DOCKER REPO TAGS UPDATED ocfedora 172.30.38.247:5000/caiqian/ocfedora latest 48 minutes ago later, someone pushed an updated image to docker.io, # docker push caiqian/ocfedora subsequent using of that image in openshift will still use the old image. $ oc new-app caiqian/ocfedora $ oc get is NAME DOCKER REPO TAGS UPDATED ocfedora 172.30.38.247:5000/caiqian/ocfedora latest 54 minutes ago $ oc new-app --docker-image=docker.io/caiqian/ocfedora --> Found Docker image 50a7e62 (About an hour old) from docker.io for "docker.io/caiqian/ocfedora" * An image stream will be created as "ocfedora:latest" that will track this image * This image will be deployed in deployment config "ocfedora" * The image does not expose any ports - if you want to load balance or send traffic to this component you will need to create a service with 'expose dc/ocfedora --port=[port]' later --> Creating resources with label app=ocfedora ... error: imagestreams "ocfedora" already exists deploymentconfig "ocfedora" created Version-Release number of selected component (if applicable): https://console.ded-int-gcp.openshift.com/ How reproducible: always
And, it seems does not allow you to change the pull spec of the image from the console as well. The change is lost after clicked "Save".
The only way I found working so far is to delete the image stream entirely.