We had a number of customers report that after upgrading their Apple Mac Pro 6,1 to the latest ESXi 6.7 Update 1 release, the system failed to boot and the following error was seen on the console:
Shutting down firmware services...
Mutiboot buffer is too small.
Unrecoverable error
A few customers made the observation that this was in result of a newer version of the Apple Firmware that was included in the Mac Pro 6,1 systems which seemed to have cause this problem. In addition, this also impacts brand new installations of ESXi 6.7 Update 1 on new Mac Pro 6,1 that includes a newer version of the Firmware. At this time, is it NOT recommended to upgrade to ESXi 6.7 Update 1 if you are running the Mac Pro 6,1. A request has already been made to the VMware HCL team to remove Mac Pro 6,1 for ESXi 6.7 Update 1 and hopefully should be reflected by later this evening.
UPDATE 04/11/19 - This issue has been resolved in vSphere 6.7 Update 2
Note: This issue can also affect the Apple Mac Mini, which is not an officially supported platform for ESXi.
In the mean time, VMware has published KB 59660 which provides a workaround that may help customers who have upgraded their Mac Pro 6,1 to ESXi 6.7 Update 1 to rollback to the previously installed version of ESXi. Please refer to the KB for more details and you can subscribe to the article for future updates regarding this issue.
I did run into this several times and also recommended sticking to 6.7 GA.
When installing ESXi 6.7 U1 on a Mac Pro with the latest firmware, using 'preferVmklinux=true' method seems to result in a successful install.
Ugghh ESXi on Laptop?!?
I'll stick with VMware Fusion 10 for the Mac!
Was just re-installing my two mac minis with 6.7U1 and ran into the same issue. However after doing a PRAM reset and shutting down the mac it all started to work. On the other mac mini though I had to reboot a few times and eventually it fixed itself again. One thing to note... When it works correctly it says the following:
"Shutting down firmware services...
Using 'simple offset' UEFI RTS mapping policy
Relocating modules and startup the kernel..."
After finishing the fresh installation. I havent had any issues as yet when rebooting. Seems to start up fine every time now.
Confirmed.
Tested the PRAM reset (3 additional startup chimes) + shutdown on a MacMini5,3 after it got stuck when I applied the 6.7U1 update. At the next boot, it proceeded past the point where the error messages appeared before. It made one additional reboot (maybe to complete version upgrade), but has booted stable afterwards. ESXi build is now at 8169922. The Mac firmware has not been upgraded since it ran OSX 10.10.
FWIW, I tried resetting the PRAM and SMC on a MacPro6,1 and installing update 1 again with no difference in result.
I was able to boot with a brand new install on the MacPro6,1 to 6.7 U1. I updated the bootrom to 129.0.0.0.0 by installing Mojave 10.14.2 to the local SSD and updated to 10.14.4 to get the latest bootrom. https://www.dropbox.com/s/9wlb4sefcmlyzao/BOOT%20ROM%20AFTER%2010.14.4.png?dl=0
Removed Serial Number from image: https://www.dropbox.com/s/g8f6lxydvc858bm/BOOT%20ROM%20AFTER%2010.14.4%20-%20open.png?dl=0
Can we get 6.7 u1 added back to the HCL since it now works with latest bootrom bundled with 10.14.4 of MacOs Mojave?
Hi Gokalp,
Thanks for sharing the update regarding the latest bootrom from Apple. At this time, the fix that's been put in place will be included in the upcoming vSphere 6.7 Update 2 release (which was just announced today, see https://blogs.vmware.com/vsphere/2019/04/announcing-vsphere-6-7-update-2-vsphere-platinum-updates-and-vsphere-robo-enterprise.html) and will be available at a later time. Once ESXi 6.7 Update 2 is released, the HCL will be updated to add support for the Mac Pro 6,1 and there is no plan to update that for 6.7 Update 1 at this time
Just hold down the alt key to force the mac boot device Selektion. Then press and hold down STRG while klicking on the Boot drive icon (the arrow should turn into an Cycle Arrow). After this ESXi is Booting Fine every Time.
Any idea if its normal for the Macpro 6,1 running esxi 6.7 u3, to go to sleep (running but not responding to ping) after the display is disconnected ? Reconnecting the display does not bring it back.
Update: The Macpro is not asleep. It continues to function correctly after the display is removed, however as soon as I plug in the thunderbolt it crashes immediately. This seems to point to a driver issue. Not sure how to figure this out, but if I do I will post it here.