I have High Sierra (10.13.3) host OS running Fusion 10.1.1 on a late 2011 MBP.
Whilst updating a guest VM running High Sierra 10.13 to 10.13.3... the update downloads fine, and then the update's installation process initiates a reboot of the VM. Upon restarting, the progress bar advances, the VM reboots again, the progress bar advances again, and eventually the progress bar disappears and all I see is a dark screen with an white apple logo. At this point the restart process is hung (I've left it sit like this for many hours). The host OS shows that the VM is using 100% of one CPU core (apparently in a tight loop).
To eliminate other variables (like 3rd party software, drivers, etc) I've created a new/fresh High Sierra 10.13 guest VM (Just an OS, no other software in the VM). I snapshot'd the VM before proceeding to the installation of the 10.13.3 update, so that I could run a number of different test permutations, reverting back to the snap shot between each test.
After numerous tests I've observed that;
- If I configure the VM with only 1 vCPU, the update installation completes successfully every time.
- If I configure the VM with 4 vCPU's the update installation process will hang (as described above).
- If I configure the VM with 4 vCPU's, and nvram boot-args="-v", the restart will sometimes complete, but still sometimes hangs.
The symptoms are similar to those reported by cragsdale1 in this thread but without FileVault in the guest VM.
Good news - this is easily and 100% re-creatable.
This would seem to suggest some sort of race (or deadlock) condition in Apple's update logic for 10.13.3.
Or, maybe in the VMware hypervisor?
Wondering if anyone else is seeing this? Thoughts anyone?