[feature] retries for copy operations

[gnoling]

Is this RFE related to an Existing Problem? If so, please describe:

• In unstable networks, there is no retry mechanism when pushing or pulling images.

Describe Proposed Solution(s):

• They would like the ability to set retries within Hauler (image level retries, not manifest level retries). Something like --retries 3 for when an image fails to pull/push.

Describe Possible Alternatives:

• Currently the work-around is looping hauler until it completes successfully.

Additional Context:

• There are times when customers are down range and networks are unstable when pulling/pushing images.

[amartin120]

Currently hauler store sync will retry things up to 3 times. It's just currently not flexible as to the "number" of retries. That's probably something that could be done though. I think the original thought was, if we tried 3 times and it still fails... it's just not going.

hauler store copy is currently the exception as it just passes a directory to cosign load and says... go do it. This will take some effort in our cosign fork if we want to include retries there.

[zackbradys]

partially similar to Issue #323...