macOS Sierra Update 10.12.4 issue and fix

bryan_feuling
New Contributor III

We have noticed an issue with the new 10.12.4 update and 13" MacBook Pro (w/ touchbar). In order to block this update, without blocking the App Store, we are pushing the following script in a policy:

#!/bin/bash

sudo softwareupdate --ignore "macOS"
sudo softwareupdate --ignore "macOS Sierra"
sudo softwareupdate --ignore "macOS Sierra Update"
sudo softwareupdate --ignore "macOS Sierra Update-10.12.4"

If you need to allow the users to update, after testing, running the following script will reset the software update to allow this:

#!/bin/bash

sudo softwareupdate --reset-ignored
59 REPLIES 59

Bhughes
Contributor
FYI: the Tripp-Lite U444-06N-DGU-C adapter will also cause this

@grepoli Which issue will it cause?

peguesrc
New Contributor

Beta still has not provided a fix. Tried updating to Beta and the same issue persist. Spoke with our TAM and let them know. The weird thing about this issue is that various workarounds work but none 100%.

jhalvorson
Valued Contributor

A few MacBookPro13,3 were being dropped off to me because the techs could not reimage or get a clean OS installed on them. They were behaving oddly after updating to 10.12.4, so one of the things they tried was to reimage them.

Reseting the SMC, zapping the PRAM, or using different types of Ethernet adapters did not resolve future year issue. The one's I've worked on have the year set to Dec 2040 or Jan 2041.

I've been able to (temporarily) resolve by booting from a USB drive with the 10.12.4 installer, opening Terminal, and then using the Date command to reset the hardware clock. Once set properly, then it will successfully netboot, net image, and/or Internet Recovery, or install 10.12.4 from a external USB drive.

Has anyone noticed that it's computers that are upgraded from 10.12.x to 10.12.4 via software updates that are effected - something included in the update process creates the problem? I wonder if installing 10.12.4 on a clean device will avoid the issue from coming back.

Using the date command gets them going again, but might not be the long term solution.

stammbach
New Contributor

@jhalvorson It also happens on a clean 10.12.4 installation.

jhalvorson
Valued Contributor

@sstammbach Yep. Since my prior post, I've witnessed that as well. AppleCare Support has informed me this will be addressed with 10.12.5. They didn't comment if it was an issue specific to which type of network adapter as a trigger. I'll see if I can get repeated problems with certain adapters or not.

stammbach
New Contributor

@jhalvorson We only had issues with the Belkin USB-C Adaptor so far. No Problems with Dell, LMP, etc.
Yes, 10.12.5 (beta 4 & beta 5) are working fine until now. Hope it stays that way with the final release of 10.12.5...

jhalvorson
Valued Contributor

Limited testing, I can get the date to jump to 203x or 204x periodically when: rebooting with a White Belkin USB-C to Ethernet (F2CU040) and a White Kanex USB-C to Ethernet (KU3CGBT) connected to a wired network.

I have not seen the date jump happen when using a Thunderbolt 3 to 2 PLUS Apple TB2 to Ethernet adapter. Nor when using a Aluminum StarTech USB-C to Ethernet (US1GC30A).

mbezzo
Contributor III

I'm definitely seeing this on non late 2016 MBP's too. Opened a case with Apple and their suggestions aren't working on this specific machine. Turning FV off allows the Mac to boot, but as soon as it is re-enabled back to boot looping.

Fun.

YoshiiZee
New Contributor II

We are having this issue too. Worst possible time as well since we've just begun the JAMF rollout to replace our current solution of Centrify.

In any case we've had about 3-4 users affected. We initially thought it was JAMF causing it, based on how we did the migration but have now seen it happen on out of box models. It appears that FV 10.12.4 Belkin USB C to Gigabit E are the wild cards. Problem is now our machines are stuck with FV enabled and unable to disable or run the following which JAMF provided us:

  1. Boot into single user mode and run the below commands

mount -uw /
kextcache -u /
exit

  1. This will boot the system into the OS where you can disable FileVault, fix the time, remove any attached dongles, then re-enable FileVault.

chris_hansen
Contributor

10.12.5 released. "...Resolves an issue, affecting some enterprise and education customers, that may cause the system date to be set to the year 2040...." among other things.

YoshiiZee
New Contributor II

@chris.hansen fingers crossed running on one of my plagued machines! I was able to finally decrypt, but odd procedure. Ran the decryption first in Recovery Mode but it kept saying Paused. Booted into Target Disk Mode and then ran the command to check the status and sure enough it was chugging along. Thought it was sort of bizarre that the decryption was actively running in TDM. 3 hours later and I was back into the users' machine.

Running the 10.12.5 patch, re-encrypting and praying for the best.

peguesrc
New Contributor

Has anyone had any luck updating from 10.12.3 to 10.12.5 without the device being stuck in the reboot loop?

mbezzo
Contributor III

10.12.5 has fixed the issue for us. I booted from an external bootable drive and installed the 10.12.5 combo updater. The computer to a LONG time at the boot progress screen, but eventually logged in as the user just fine.

peguesrc
New Contributor

@mbezzo When you updated to 10.12.5 were you computers on 10.12.3? Can you elaborate on combo updater please?

peguesrc
New Contributor

@mbezzo answered my own question, nvm.

Not applicable

2016 15" MBP w/ TB worked fine for 3 months before 10.12.4. After a clean install of either 10.12.4 or even 10.12.5, it still will enter a startup loop after a few normal restarts. I've resorted to using 10.12.2 for now until a definitive fix arrives.

I always have FileVault turned on.
I don't use any Belkin adapter.
I've never noticed an issue with the date going to the year 2040, etc.

AppleCare knows about the issue but wouldn't provide their in-house tracking number regarding it; don't think it's a Campaign as of yet, since they have no published fix so far.

knnuui
New Contributor

Hi everyone, it sucks to have an update causing the boot loop issue, but good to see that I'm not alone. From my understanding jamf is a website for IT professionals but I'm not one. I saw mbezzo mentioned installing 10.12.5 from an external bootable drive to fix the issue. I would like to know if anyone has luck fixing the 10.12.4 boot loop by installing 10.12.5 update from the App Store? Thanks!

chriscollins
Valued Contributor

@knnuui the only thing that worked consistently for me was the same thing suggested by @YoshiiZee

1.) Boot into single user mode
Run the commands:
mount -uw /
kextcache -u /
exit

2.) As it continues to boot, it will look like it's stuck but it isn't. Just wait a while and it will eventually reach the login

3.) Disable File Vault encryption and let it completely decrypt

4.) Re-enable FV

Sometimes booting to something else and fixing time would work for a bit but not always. Also booting from an external drive and trying to install the combo update for 10.12.5 to the regular boot drive would not work. Once the update had finished running it would just be back into the reboot cycle.

JefferyAnderson
Contributor

10.12.5 update worked on my first test machine so far. I used the App Store update to go from 10.12.3 to 10.12.5. This one machine I had tried to go to 10.12.4 but kept getting the time bug, so I wiped and reinstalled 10.12.3 and kept it at that version until today. I'm glad to have this issue resolved.

knnuui
New Contributor

@Jeffery_Anderson Good to know!