I have noted that sometimes the entire policy history gets flushed for a machine unintentionally. In some instances, a change in the machine name instigated an entire policy flush, and that machine started getting all the policies all over again. Is this normal and expected behavior?
I have seen a simple recon cause this as well. I recon'd a group of machines one day last week and they all kicked off policies that had already ran weeks before. Their history was clear.
Does a simple recon mean that the jamf binary/policy was used to update inventory or does a simple recon mean that the Recon Application/User Initiated Enrollment was used? If it was done via the jamf binary and the command line this is not the expected behavior. If this was done via the Casper Recon Application then the behavior that you are seeing is the intended behavior. In version 8.2x the jamf binary spilt the recon verb into recon and enroll. The recon verb is now only used for updating a computers inventory where the enroll verb enrolls a computer into the JSS. Any time that a machine enrolls to the JSS it should flush policy history as the JSS views it as a new machine. Hopefully this information helps.
Joel, Yeah, I learned that the hard way… I used the Recon app to scan a network segment and bang!, all systems in that segment got re-enrolled. Maybe a feature request… to allow Recon to actually recon systems as well as enroll.
