Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 Mac OS 11.3 has broken support for older Mac Pros
- Thread starter VitaminK
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
- Status
Why not. Engineers tend to support stuff in their spare time because they like. Seen this quite a few times in my life.
If it was by design, they wouldn’t let it crash arbitrarily. Quite hard to specify as feature. They may not care, but if this was on purpose, it would be way more deterministic and stop with a simple error message…
Even if Apple engineers know how to fix the issue, why should they endanger their position with Apple and breach their NDA's just to help with the debugging or provide us with a solution?
Yes, I agree. There seems to be no real pattern emerging from my tests as far as a successful method among the variations I tried. But besides methods I tested, the booting froze almost always at the same point in the verbose -- with only one exception I believe. And this is why I wanted to include several verbose pics, even if they are almost identical from each other. I guess this is the only real pattern emerging from my relatively limited tests, just like others (like @JohnD) found out before me.
There is something about this. It seems that on a failed boot, the system loses access to storage devices early in the rooting phase. This occurs even before we see the hang in verbose mode. To test this hypothesis, I booted macOS on a supported Mac from a USB drive, and unplugged the drive at around #[EB|LOG:EXITBS:START]. This provoked a hang later in the process similar to what we have been observing. Of course, this is on a different machine, so the hanging doesn't occur at the usual ethernet initialization point. It would be interesting to repeat the experiment on a Mac Pro 5,1.
Did the debugger reveal anything that would support the hypothesis of losing access to storage devices?
I observed the same thing when I tried to boot my cMP from an external USB SSD. The case has a combined power/activity led.
If you try to boot without verbose mode, then the mac turns off the external ssd before it hangs.
The led is off and I know immediately that my mac is not booting.
However, when I use verbose mode, the behavior is different and the drive's led is not turned off.
If you try to boot without verbose mode, then the mac turns off the external ssd before it hangs.
The led is off and I know immediately that my mac is not booting.
However, when I use verbose mode, the behavior is different and the drive's led is not turned off.
Yes, the NMVe LED light was very important to understand what was happening and when it was happening. You are right @cdf -- if the LED does not show activity right away at boot, then it's a failed boot, regardless of the status bar activity (in my case status bar usually goes up to about 1/4 even when there's no NMVe LED activity).
The challenge in my case was/is that my NMVe PCIe card has no activity LEDs!!! Luckily I had some LEDs and old cables around and improvised a temp activity LED for the boot NMVe drive. It worked fine and helped me a lot to understand the timing of fails. For your entertainment, I attach some pics of my "professional" LED activity setup extending all the way to the table top, just next to my mouse... ?
The challenge in my case was/is that my NMVe PCIe card has no activity LEDs!!! Luckily I had some LEDs and old cables around and improvised a temp activity LED for the boot NMVe drive. It worked fine and helped me a lot to understand the timing of fails. For your entertainment, I attach some pics of my "professional" LED activity setup extending all the way to the table top, just next to my mouse... ?
Attachments
Well... I have upgraded two 5,1 today to 11.4 and here's what happened (copied from another thread / OpenCore):
Strange things happen... One of my two 5,1 did perfectly fine with the system preferences update (3GB download or so) & the whole thing took maybe 15 minutes. No issues.
The other machine kept canceling out of the update & I ended up manually getting the whole 12GB full install and the update took around an hour and required multiple reboots. Both machines are running 11.4 perfectly OK now.
More info: My main machine (the one that initially didn't work) is running from an internal PCI NVM & at first I thought that might be the issue, so I plopped my original SSD with 11.3.x OC 0.6.9 into the other 5,1 and the install kept canceling out also with a download error. Something is up with the original install, definitely nothing to do with the PCI NVM. Both machines are 4,1 to 5,1 – one's a Dual 3.46 12-core with a Radeon RX 580 8GB and the other one's a 3.46 6-core with a Radeon RX 560 4GB... Both recently upgraded from Mojave.
TL;DR
5,1 12-core 3.46 PCI NVM, 64GB, Radeon 580 RX, USB 3 PCI – 11.4 update canceled with error, got it to work with full 12GB download
5,1 6-core 3.46, EVO SSD, 32GB, Radeon 560 RX – 11.4 update worked via system update
Strange things happen... One of my two 5,1 did perfectly fine with the system preferences update (3GB download or so) & the whole thing took maybe 15 minutes. No issues.
The other machine kept canceling out of the update & I ended up manually getting the whole 12GB full install and the update took around an hour and required multiple reboots. Both machines are running 11.4 perfectly OK now.
More info: My main machine (the one that initially didn't work) is running from an internal PCI NVM & at first I thought that might be the issue, so I plopped my original SSD with 11.3.x OC 0.6.9 into the other 5,1 and the install kept canceling out also with a download error. Something is up with the original install, definitely nothing to do with the PCI NVM. Both machines are 4,1 to 5,1 – one's a Dual 3.46 12-core with a Radeon RX 580 8GB and the other one's a 3.46 6-core with a Radeon RX 560 4GB... Both recently upgraded from Mojave.
TL;DR
5,1 12-core 3.46 PCI NVM, 64GB, Radeon 580 RX, USB 3 PCI – 11.4 update canceled with error, got it to work with full 12GB download
5,1 6-core 3.46, EVO SSD, 32GB, Radeon 560 RX – 11.4 update worked via system update
Here's a thought... I also experienced issues when I had 11.3.x on an SSD and a fresh PCI NVE – the machine DID NOT BOOT from either drive with the other one present. The NVE is a clone from the SSD... As soon as I took out the SSD with the same OS, everything worked fine. I have another SSD in there with Catalina and it creates ZERO issues.
Now, if I stick that Big Sur SSD in my second 5,1 with Big Sur present on its own boot SSD, no problem. strange indeed.
Now, if I stick that Big Sur SSD in my second 5,1 with Big Sur present on its own boot SSD, no problem. strange indeed.
My setup:
1) 5,1
2) Thunderbolt slot 4
3) USB 3.0 slot 3
4) NVME slot 2
5) RX560 Slot 1
I removed the USB3.0 as it was showing a number of issues with it, so I thought I would remove any errnous lines in the images that way.
Currently booted on 11.3.1 with the USB3.0 removed.
Boots were as follows:
1) during 11.2.3 login, upgraded to 11.3.1 with 0.6.9 ML opencore with Syncretic updates (booted on 11.2.3 before applying the 11.3.1 update)
2) upon reboot, system seemed to start the patch
3) next reboot, system blinked into GUI for a second then rebooted
4) next reboot, system froze / hung on the NVME (I'm assuming at this point it was on a 11.3 kernel)
5) cold boot, system continued install and seemed to function fine after that
6) Booted into and logged into 11.3.1.
7) Reboot, hang
8) reboot, hang
9) cold boot, booted again
I disabled hyperthreading (had to boot into 11.2.3 USB fob), no change, system hung on boot. At this point, I removed the USB3.0 card and booted again, failed, cold boot and was able to boot into 11.3.1 again.
I'm attaching the photos of the verbose booting in 11.3.1
1) 5,1
2) Thunderbolt slot 4
3) USB 3.0 slot 3
4) NVME slot 2
5) RX560 Slot 1
I removed the USB3.0 as it was showing a number of issues with it, so I thought I would remove any errnous lines in the images that way.
Currently booted on 11.3.1 with the USB3.0 removed.
Boots were as follows:
1) during 11.2.3 login, upgraded to 11.3.1 with 0.6.9 ML opencore with Syncretic updates (booted on 11.2.3 before applying the 11.3.1 update)
2) upon reboot, system seemed to start the patch
3) next reboot, system blinked into GUI for a second then rebooted
4) next reboot, system froze / hung on the NVME (I'm assuming at this point it was on a 11.3 kernel)
5) cold boot, system continued install and seemed to function fine after that
6) Booted into and logged into 11.3.1.
7) Reboot, hang
8) reboot, hang
9) cold boot, booted again
I disabled hyperthreading (had to boot into 11.2.3 USB fob), no change, system hung on boot. At this point, I removed the USB3.0 card and booted again, failed, cold boot and was able to boot into 11.3.1 again.
I'm attaching the photos of the verbose booting in 11.3.1
Attachments
Just so I'm clear, this issue occurs with ANY storage type on the cMP or just NVME? My other MP4,1 (converted to 5,1) has AHCI controller and disk in it, not NVME. (I was snooping on my main system and just now noticed this, I also have a spare AHCI 256Gb disk (though, I need to see if I have a compatible card or not)).
Is this a problem also if virtualisation is turned off? I had similar problems with 11.2.2 when virtualisation was turned on.
(similar in regards to the 5.1 not booting)
(similar in regards to the 5.1 not booting)
Thanks for posting your experience here. The real test is to confirm whether multiple boots (cold and warm) will work fine or fail. If you tend to restart or shutdown your machines regularly, then let us know how it goes, whether you have boot failures or not. If you read the last 2-3 pages of this thread, you will find various accounts of boot failures as reference.
Thanks.
Unfortunately, yours is consistent with previous reports of boot failures during upgrade and regular use. And your verbose, too, froze at the same spot as most previous reports.
What happens if you turn off hybridization in the OC config file? It probably has nothing to do with this, but I were unable to boot on an earlier version of Big Sur when hybridization was turned on.
The issue is still there. In fact, the issue is also there when booting Big Sur with no_compat_check and no OC (which is actually possible).
Not sure what I did right to some others but I went from 11.2.3 to 11.4 without a hitch, just like a regular upgrade on my Mac Pro 5,1 (Mid 2010), Radeon 5700XT, Highpoint 7101a - PCIE card, and Samsung 970 NVME SSD, USB 3 PCI-E card.
The only thing I did was to detach all the other drives from the sleds so only the NVME was attached.
Rebooted a couple of times and did give me a warning symbol but I did nothing the machine restarted itself anyway and am now writing this message on here via the Mac Pro.
Using OC 0.6.7.
No sleep issues, no problems with Wifi or Bluetooth. Everything seems a little bit snappier.
Just done a cold boot and warm restart. No problems. Super happy.
The only thing I did was to detach all the other drives from the sleds so only the NVME was attached.
Rebooted a couple of times and did give me a warning symbol but I did nothing the machine restarted itself anyway and am now writing this message on here via the Mac Pro.
Using OC 0.6.7.
No sleep issues, no problems with Wifi or Bluetooth. Everything seems a little bit snappier.
Just done a cold boot and warm restart. No problems. Super happy.
Sure, once booted, there are no issues. It's the booting process that's not reliable. Many of us have been able to reboot many times in a row, especially with most drives disconnected, but always eventually the boot will hang.
Perhaps the following observation is irrelevant, but, just in case, here goes: I may have missed it, but I haven't seen any reports of a supposedly successful series of boots using versions of Big Sur after 11.2.3 on a Mac Pro SERVER 5,1. I ask myself if the fact that you have the Server component installed is working wonders in your case.
- Status