I was on a shared session with rhn-support-sjavurek today and we have noticed multiple failure with the provisioning of a remote ssh container.

The error reported on Fabric shell was the usual dump of all the .sh scripts with no specific hint on the underlying problem.

Trying to perform a wget operation from the remote host to the root node, showed that even wget had to resume the download multiple times since it dropped.

We probably need to either implement a more reliable resume behavior or to identify these specific errors and report them to the user so that he might become aware of networking issues.