| Currently, we update the CA and CRL `last_updated` time before we try to actually acquire them from the server; if a problem occurs and the actual crl/ca does not successfully update, puppet will not try again until we hit the interval for updates. We initially wrote it this way to prevent possible churn in errors if we somehow errored out prior to updating/touching the files on disk. Now that the code has been proven to be more stable after release, we could change the update time to happen after successfully updating the CA/CRL. This way, if puppet does error out during a run and not update the CA/CRL successfully, it will try again on the subsequent puppet run, and not wait for the respective refresh intervals. |