I have a really odd problem that I can't get to the bottom of.
I am provisioning devices by going through the setup assistant and then using a quickadd package that then enrols the device, runs a policy and then reboots and auto logs into an account. The devices are then provisioned by running a policy via self service. Each policy in turn runs a script that runs several other policies and then restarts the device ready for use.
This all works fine, but sometimes following the reboot the mac will slowly run out of memory and eventually just restart and create a load of jetsam files under the DiagnosticReports directory.
If I sit and watch a problem device, the process "kernel_task" just keeps climbing in use but you can only see this if you enable the "real memory" column under activity monitor. Under the normal "memory" column it will say it's only using 100Mb or so memory.
I'm really at a loss as to how to work out what is causing this as the only thing I have to go on is the jetsam files that get created that suggest "kalloc.48" is the process that eventually kills it.
What's even stranger is that once it's crashed once it will never do it again so I then have to re-provision to try and replicate the issue. It's intermittent as well which is really helpful...
This issue seemed to start on its own as well, when I first started rolling out Mojave I didn't have this issue.
My Jamf Pro dev server is running 10.13 and I can't replicate the issue but my production server is still 10.12 but I don't see how anything Jamf could cause this issue.
