You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi!
I am developing some content for RHPDS, using catalog snapshots to fix the images of the operators that are used. This is standard procedure to make sure a deployment workload will work exactly in the same way along the Lab lifecycle.
Two times already in the past months I have been impacted by digests changing in Quay. As a catalog, and therefore its snapshot, uses SHA to refer to images, deployments are failing because the images cannot be found any more if the digest changes.
For example, I'm using a community-operator snapshot from 2021_07_01 (so quite recent). The subscription creates an Install plan with this reference for the ODH operator image: quay.io/openshift-community-operators/opendatahub-operator@sha256:7cfa3371c22fb360793ea1e6a1e596fa9e8493300e509d353125f27be850c3d3
Which does not exist any more as apparently all the images have been recreated 9 days ago, apparently due to this change: operator-framework/community-operators#4172
So, is it the standard procedure to completely change images in Quay sometimes, even when there is no change at all with the operator image itself? If that's the case, is there a way to avoid the issue I am facing?
I know I could bundle the operator image itself with the catalog as part of my workload, but that would become really huge if I have to do this for all operators I'm deploying as part of this workload.
Thanks!
The text was updated successfully, but these errors were encountered:
@guimou, we are aware of it. We will not overwrite bundles in the future anymore. Current setup overwrites only if contributor insists on it. We are working on a feature preserving digests. Sorry for any inconvenience.
@J0zi OK, fair enough, I'll switch back to my snapshots in that case. Then if the contributor insists in messing up everything again I guess it becomes their problem :-)
Thanks a lot!
Hi!
I am developing some content for RHPDS, using catalog snapshots to fix the images of the operators that are used. This is standard procedure to make sure a deployment workload will work exactly in the same way along the Lab lifecycle.
Two times already in the past months I have been impacted by digests changing in Quay. As a catalog, and therefore its snapshot, uses SHA to refer to images, deployments are failing because the images cannot be found any more if the digest changes.
For example, I'm using a community-operator snapshot from 2021_07_01 (so quite recent). The subscription creates an Install plan with this reference for the ODH operator image: quay.io/openshift-community-operators/opendatahub-operator@sha256:7cfa3371c22fb360793ea1e6a1e596fa9e8493300e509d353125f27be850c3d3
Which does not exist any more as apparently all the images have been recreated 9 days ago, apparently due to this change: operator-framework/community-operators#4172
So, is it the standard procedure to completely change images in Quay sometimes, even when there is no change at all with the operator image itself? If that's the case, is there a way to avoid the issue I am facing?
I know I could bundle the operator image itself with the catalog as part of my workload, but that would become really huge if I have to do this for all operators I'm deploying as part of this workload.
Thanks!
The text was updated successfully, but these errors were encountered: