feature request: add delay between --retry-times

[xwu2intel]

This is my calling sequence:

ssh -4 <user>@<ip> -O ControlMaster=auto -fNL localhost:13340:localhost:20048     # localhost:20048 is the remote registry
skopeo copy --retry-times 10 --dest-tls-verify=false docker://<image> docker://localhost:13340/<image>
ssh -4 <user>@<ip> -O cancel -L localhost:13340:localhost:20048

As the remote registry is not directly accessible, I have to use ssh port forwarding as a bridge.

This works fine if there is no error during skopeo copy. If there is any error in the first try, then any subsequent retries will fail as well. This is because the remote registry is not immediately responding to subsequent connections, until after a time delay.

This following sequence works fine in the case of retries:

ssh -4 <user>@<ip> -O ControlMaster=auto -fNL localhost:13340:localhost:20048     # localhost:20048 is the remote registry
for retries in { 1..10 }; do
    skopeo copy --dest-tls-verify=false docker://<image> docker://<image> && break
    sleep 5s
done
ssh -4 <user>@<ip> -O cancel -L localhost:13340:localhost:20048

The request is to add time delay between retries.

[mtrmac]

Thanks for reaching out.

c/common/pkg/retry already implements a Delay option, so we just need to hook that up in the CLI.