Imaging Failures (Mac OS X 10.11.6, JSS 9.93)

georgecm12
Contributor III

Our machines are iMac14,3 computers with 8GB RAM and 250GB Apple SSDs.

We're following our normal lab fall reimaging protocol by net booting our machines (using a disk created with AutoCasperNBI 1.3.3), then executing a Casper Imaging configuration that contains an AutoDMG-generated 10.11.6 disk image and our set of packages.

We're seeing about a 60% failure rate. We see three distinct failure states:
1) Some machines boot to the Temporary Adobe Account and just... stop. It appears they don't even bother to run the Casper Imaging-generated postinstall.sh script. Rebooting the machine does not seem to help. 2) Other machines boot to the Mac OS X Setup Assistant.
2a) On one machine, I tried booting to the recovery disk, doing a touch /var/db/.AppleSetupDone, then rebooting, but they still boot to the setup assistant. Still a little confused by that one.
2b) On another machine, I walked through the Setup Assistant, creating an admin account, It appears, from the /Library/Logs/JAMF/ImagingScripts.log, that on this machine the postinstall.sh script executed previously... but once I went through the Setup Assistant and created another account, it kicked off the script again.
3) Still other machines enter a constant reboot loop. They don't appear to be panicking, they simply boot to the Apple logo...then when it would normally drop into Finder, it reboots.

The odd part is that other machines - identical in every way, in the same room, on the same network segment, complete just fine.

Right now, my only solution is to start those machines that fail over again. I'm at a bit of a loss. Any ideas?

7 REPLIES 7

tthurman
Contributor III

Do the machines that fail, fail every time?

For the kicks, can you wipe the machine completely out from Internet Recovery -> Disk Utility and try imaging it again? (In the past - I've seen some weird stuff with the base install of OSX.)

georgecm12
Contributor III

@tthurman Nope. I was, eventually, able to get the lab I was working on reimaged. I will have to see how the next lab goes when that gets started this week.

I do an erase then restore from image in Casper Imaging, prior to applying the packages. That should accomplish the same thing.

lakem
New Contributor II

Bump this one. I am having the same issue in a number of classrooms I am re-imaging. It is starting to royally p**s me off whenever I see the desktop of the temporary Adobe install account, or the Apple setup assistant, instead of the Casper Imaging "installing software..." screen. Let me know if you have found any solutions to this, otherwise, I may start using Deploy Studio again.

thoule
Valued Contributor II

When you see the Desktop and Temp Adobe Account, open Console and watch /var/log/jamf.log. You can see what/if JAMF is doing anything there. Did you upgrade your JSS recently? And netboot image still have old version of Casper Imaging on it...?

mojo21221
Contributor II

Im having the same issues in my org. Issue didn't start happening till 9.93 and 10.11.6. Verified the netboot image is running 10.11.6 with 9.93 imaging. Since they both came about around the same time. I can't determine nor have had the time to figure out which is the cause. Hoping for a fix in 9.96... Anyway, just wanted to say you are not alone.

bmarks
Contributor II

@georgecm12 For the ones that get stuck in the Adobe user, I have noticed that if I turn off bluetooth and restart, then the next step (curtain mode) does complete successfully. Maybe worth a shot. However, I don't know if my root cause is the same. The reason I started turning off bluetooth was because I had one that was different than all the others that would get stuck in the Adobe user and pop up an alert about the BT PAN connection. In subsequent tests, my solution ended up working on others that also got stuck in the Adobe user but with no popup.

mcrispin
Contributor II

Bump. Possibly same thing here, affecting c. 30% of all imaging.

We are using 9.91 and 10.11.6, building images via a thin process on top of a AutoDMG 10.11.6 set. We can duplicate using DeployStudio and a simple restore process. For those who are having this problem, have any updates? I would be very curious to know what the file system of a damaged machine looks like. We are seeing a "-pv" folder on the root, and the symlinks broken with a visible "var" folder sitting on root. The problem we are seeing also erases the /Library/Logs/ folder. Our fleet is also mostly iMac14,1 and 14,3.