Bug in PostHTTP triggered by DNS issues

After a successful POST, if doing 2-phase commit, we loop indefinitely until we get a response from the server or the user stops the processors. If DNS issues pop up or a firewall, we instead get IOException, and we loop indefinitely, destroying system resources.

This issue is the result of continually trying to delete the HOLD until we successfully receive an HTTP Response from the Server. We should instead change the logic to retry only some configurable number of times and wait 1 second in between retries.