Skip to main content

I have a policy to cache Mavericks to a scoped client. The policy excludes all clients that belong to a smart group of machines that have the "Install OS X Mavericks.InstallESD.dmg" already cached. When the dmg is cached it runs an inventory update to update the smart groups.



The problem that i've been running in to is that when the policy fails due to a loss of connection with the JSS or whatever, i have to go in and flush the error log to have it run again. So, what i want to do is set the policy to "ongoing" so when it does fail, it will just resume when it checks in again.



My question is, if the policy is set to ongoing and to run at checkin(every 15 minutes), will multiple sessions of the same policy run if the download takes longer than 15 minutes? my assumption is no and that this policy will only run again if it fails to cache the installer.

I'm not 100% sure here, but I believe that, while you can potentially have policies from different triggers running concurrently, ones based on the same trigger should not kick in until the current policy run is complete. Meaning, if you'use the every15 check-in trigger, even though its set to Ongoing frequency, it should not start a second instance based on the every15 trigger while one is already happening.



Maybe someone can double check me on that, or perhaps JAMF can chime in. But I do think that's how it works. At least, at times when I've tried to manually run a jamf -trigger every15, I sometimes get a message that the trigger is already running and needs to complete before being called again.

Hi,



If I use the check-in trigger, and the policy fails (as in example above) and the policy is set to run once for the computer will it ever run again? I would have expected it to keep retying every check in until successful...?



Thanks

it will not. the policy runs once, it does not succeed once. I believe this is a feature request.... ahhh here it is: https://jamfnation.jamfsoftware.com/featureRequest.html?id=827

Yeah, JAMF really needs to change that behavior to be like run once per computer until success. Because how it works now doesn't make any sense. "Failed" ? "Ran" in my book.

Indeed that is a bit of a pain, I guess if I am deploying an application I'll need to set the frequency to once an day then use a smart group which excludes machines where it has succeeded, based on the package being installed. That way it should fall out of the group once it has succeeded.

Terms & ConditionsCookie settings