10.9.3 - Post-imaging hang at login window

jreinstedler
New Contributor III

Seeing an issue here with 10.9.3 - after imaging a machine with a very minimally configured base image and nothing else in terms of packages, the machine hangs at the login window. I built the image yesterday and had the problem, then built a new 10.9.2 image with the same exact config resulting in no issues. Then built the 10.9.3 image again today and saw the issues return.

Eventually I am passed through to the desktop, but only after waiting around 5 minutes. Will get beach-balled off and on the whole time - credentials can be entered, however they're slow to show due to the hang. Rebooting after reaching the desktop is as fast as it normally is and does not result in a hang.

Checked the logs and found com.apple.launchd logging an event repeatedly until the hang ended. Specifically, "(com.apple.cfprefsd.xpc.daemon): Throttling respawn: Will start in 10 seconds" - 10 secs pass in the logs and then it rolls again...

Am also seeing another entry mixed in with the last one... not as often, but still fairly often: "xpcd[24]: Info.plist does not contain an XPCService dictionary: /System/Library/Frameworks/Security.framework/XPCServices/SecurityAgent.xpc"

Image was captured via Composer from a MBP Retina 15" in TDM. Netboot server on 10.9.3 & still running 9.31 apps & JSS.

9 REPLIES 9

bentoms
Release Candidate Programs Tester

Create the image via AutoDMG from the 10.9.3 installer.

mibrodt
New Contributor

I have been seeing similar behavior. I have tried using a thick image and a thin image workflow, both have the same result. I am able to log in eventually, but then I get pinwheeling ever couple of seconds. It is the com.apple.cfprefsd.xpc.daemon, or com.apple.cfprefsd.xpc.agent crashing and then re-spawning... I've yet to find a solution to it, but I'll try the AutoDMG to see if that helps.

jreinstedler
New Contributor III

In regards to workflow - I am using as thin as it gets. The OS image + Recovery HD and nothing else is in the workflow (I did not add anything initially as I was planning on testing 10.9.3 deployment first). The OS image is as close to bare as possible. I have it configured for one admin account (so we can login immediately after reimaging) and the recent updates that dropped post-10.9.3 release. I'll consider doing this a different way later (with built-in accounts, etc), however for now, I need it to work the way that it was designed to and without the aid of 3rd party tools.

I'll be opening a ticket with JAMF later today on this. Especially since the traditional way of doing things is known to be affected, yet the use of a 3rd party tool makes it potentially trouble free from the sound of things...

bentoms
Release Candidate Programs Tester

jreinstedler
New Contributor III

Thanks for the idea. However, as I said before, this will not work for our environment at this time.

I have a ticket opened with JAMF - will update as I get them (if they don't first...)

bentoms
Release Candidate Programs Tester

What other settings are you deploying, have you tried a no-imaging workflow?

andysemak
Contributor

@jreinstedler Did you ever get a response from JAMF regarding this issue? I'm seeing it on a few 10.9.4 machines.

jkunkel
New Contributor II

I'm also curious if JAMF has responded to this before I go any further. I need to test more tomorrow, but I found logging in as local admin generates the hang, but logging in as the hidden JAMF admin user, getting prompts for Apple registration & iCloud accounts, then logging out allowed the admin user to login rapidly on the next attempt.

jreinstedler
New Contributor III

@andysemak @jkunkel - Did not get an answer on this yet. I was out at one point and the case was closed out with no definitive resolution. I am back to working on Casper stuff again and am still running into the issue, even with 10.9.5 and CS 9.51. I have reopened my case and will see if we can get through it this time.