The reproducer steps as defined are conflicting. If I want to run remote commands on the client, I have to enable them using 'rhn-actions-control'. So, I need to install 'rhncfg-actions' package. rhncfg and rhncfg-client packages will be installed as dependencies, so they are crucial to execute the remove commands on the client. I simply cannot remove /var/spool/rhn directory, as it's a necessary part of rhncfg: # rpm -qf /var/spool/rhn rhncfg-5.10.74-10.el6sat.noarch This is not a valid scenario, we would support. I'm aware the customer would like to see the action being failed on the server (immediately). There are other mechanisms to mark the action as failed, in case it was picked, but not finished within some time. I see the action has been marked as failed after 10 times it was picked up and not finished. With osad on, it took 1-2 mins. Very low prio bug.
Even it is a low priority bug, is there any update about it? We are able to make a prediction for the customer?
I'm closing the bug as deleting directories the client tools rely on makes the customer setup unsupported. We have no intentions to do any changes in this area. Closing as WONTFIX.