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
For 0.2 we are only handling modify events where the pod is marked for termination. We need to implement functionality for stopping the old "container" and starting a new one when spec.containers[*].image or spec.initContainers[*].image change
The text was updated successfully, but these errors were encountered:
Is there a recommended way to find out whether the image has changed or not.
Unfortunately, .image holds the new image immediately after the modification. That makes it hard to find out whether the image has changed.
Would it be a good solution to introduce a new pod field, which holds the current images running in the pod?
Yeah, I wasn't sure on the best way to do this. client-go on the update method passes the old and the new object, but the kube crate doesn't. I think the ideal option is to add support for that to the kube crate, but that isn't a requirement for this feature.
I think we should probably just keep an in-memory map of any pods and their current images for now. A stretch goal would be to only keep track of bare pods (as other pods created from controllers just delete the pods and create new ones)
For 0.2 we are only handling modify events where the pod is marked for termination. We need to implement functionality for stopping the old "container" and starting a new one when
spec.containers[*].image
orspec.initContainers[*].image
changeThe text was updated successfully, but these errors were encountered: