`podman manifest push` missing retry logic that is present in `podman push`

[Jared-Sprague]

Hello!

When I podman push a large image to a registry with a short timeout, I get an EOF error when the registry times out, then the retry logic kicks in and tries again retrying in 1s ... (1/3) this succeeds because the previous layers were copied and the push finishes within the timeout.

However this is not the case for podman manifest push, which is missing the retry logic that is present in podman push.

This issue is to add the same retry logic to podman manifest push so it behaves the same way as podman push on network error.

Steps to repoduce:

1. Force a failure on copying a layer when trying to podman manifest push, such as pushing a large image to a registry with a short timeout.

Expected result:

• podman manifest push will retry 3 times before giving up, like podman push does

Actual result:

• podman manifest push does not retry and exits after the first error.

[flouthoc]

I'll check this thanks for creating the issue.