Steps to fix a catalog image from a bad umask Pull down the catalog image podman pull myregistry.com:8443/redhat/certified-operator-index:v4.18 Start up the image and get a bash shell podman run -ti --entrypoint bash --name newcatalog myregistry.com:8443/redhat/certified-operator-index:v4.18 inside the container, fix permissions like commit the new image with updated entrypoint and opm command to ignore integrity because we changed the file permissions podman commit --change ENTRYPOINT=/bin/opm --change 'CMD ["serve", "--cache-enforce-integrity=false", "/configs", "--cache-dir=/tmp/cache"]' newcatalog myregistry.com:8443/redhat/certified-operator-index:v4.18 Push the new image into your registry.
Note, you could push this to another tag to preserve the old one
podman push myregistry.com:8443/redhat/certified-operator-index:v4.18 Update the catalog source to poll for new image faster. Default is 15 minutes if not specified oc edit CatalogSource <cs-certified-operator-index>
add in the update interval like
SourceType: grpc
updateStrategy:
registryPoll:
interval: 1m0s
Below here is likely not needed!!!!!! Delete the old catalog image from the cluster by deleting the yaml for it Update the catalog yaml file to point to your new container image
If you push using the same tag, OCP may not pull your new image because it thinks t
hat it has it already.
Change the catalog yaml to use the digest of the image instead of tag unless you us
e a different tag
To get the digest, go into the quay web console and find the image and then tag
Quay has a button "pull this image by digest with podman"
Take that image and digest, put it into your catalog yaml file, then oc create that
into the cluster.
