In this scenario DSA is not configured, but there are top-level relays and a bunch of second-level relays. All of the clients are communicating with the relays, only the top-level relays communicate with the BigFix server.
Let’s say that on Monday, I create a baseline and schedule it to be deployed on Saturday at 6 AM.
At some point before that patch window starts the primary BigFix server goes down for some reason (network failure, power failure, whatever).
Will the clients still start to patch on Saturday at 6 AM or does the primary BigFix server need to be operational to tell those clients “it’s time to patch”?
My understanding is that a proportion of the clients will try to execute (i.e. those that learned of the new action before the server fell over).
If the patch has already been downloaded via the relay (or the relay is configured to get its own content from the Internet) the clients will work, otherwise they will get stuck waiting for the download to become available.
If the baselines and actions were created on Monday and the server goes away on Thursday, as long as the clients had checked in sometime between Monday and Thursday they would be able to patch themselves?
That should work fine in most cases, as long as you checked the “Begin downloads before constraints are satisfied” checkbox in the Action Settings dialog.
One exception could be actions that execute downloads using ‘download now’ or ‘download now as’. Those may depend on some logic in the running action before figuring out which downloads to request and might not be able to precache them. Another exception might be Red Hat patching, in that the RHSM Download Plugin needs to execute a command against the RPM package in order to determine prerequisite packages and may download those prerequisites at execution time.