M4 mini usb-c disconnects - a possible solution?

[DHA]

Have had a bas m4 mini since launch. Two usb c SSDs attached (one for external and one for time machine).

I have been having the normal drive not ejected errors during sleep for both drives intermittently, however never at the same time. My photos library also sits on the external drive and it would also stop iCloud sync after approx 24 hours - the only way to get sync back working was a reboot.

I had been running without put hard disks to sleep when possible, but this didn't prevent the problem. The latest option has been to disable the screen saver, set "turn display off when inactive" on, and set "prevent automatic sleeping when the display is off" on.

So far I have ran for 48 hours without either drive disconnecting. Yes it's not an ideal situation as the mini will never sleep but it has solved the issues of drive disconnects. I don't have any other usb peripherals connected so not sure it will fix that issue that others are having with keyboards etc.

 

[Boyd01]

Have had an m4 Mini for about two months. It runs 24/7 as a media server for my large iTunes library on a 4tb external USB SSD. No unexpected ejects of that drive yet. I have it set to never sleep, no screen saver, with the display sleeping after 3 hours, FWIW.

There are several threads about unexpected disk ejects, apparently the problem goes back to Mojave, if not farther. Here's a recent thread

https://forums.macrumors.com/threads/unexpected-disk-ejects-of-external-ssd.2438652/

 

[HobeSoundDarryl]

Broad and common problem. Not every drive/enclosure "unexpected ejects" but many do. This is not a brand thing, age of drive thing, etc. I've thoroughly tested this and tried every "figured it out" solution I could find on the web before finally concluding it is BUG(s) in macOS that can only be resolved by Apple debugging macOS. But I wish I had a dollar for all of the "figured it out" posts... eventually recanted when their drives unexpectedly eject again.

My very best (wild) guess is that the ports basically "crash" from time to time and some enclosures handle this "reboot" better than others. Sleep gets much scrutiny but I think that's coincidental... because there's a LOT of time while Macs are generally asleep. Analogy: while I was asleep last night, a LOT of good and bad things happened in the world. However, just because I was asleep at the time, doesn't mean my sleep caused any of it.

If not port crashes, my next gut guess is this is in power management code, likely inherited from a-series code where the focus on maximizing battery life probably attempts to work down power usage of externals to a minimum... and it just "trims" too far until finally some drives "lose connection" (and unexpectedly eject). Generally a-series stuff is not going to have long-term connections to external drives, so this code probably works fine there because user ejects before this gets to that level. But on computers- many without batteries- this code MAY be the actual cause/catalyst/contributor.

When we wake our Macs and find something is different/changed/ejected, it seems obvious that sleep caused it. But perhaps it just happened while Mac was asleep... in all of that block of time we let Mac sleep.

I can confirm the following first and second hand: I've had many "unexpected ejections" while Mac was actively writing to proven reliable drives attached to it... such as while using the external as a scratch drive for FCPX work. Obviously, BOTH Mac and Drive have to be awake to be writing files from one to the other. Unhook that same cable to that same drive and plug it into an older Mac running macOS before Big Sur (or any PC) and it is perfectly stable again. This is relentlessly repeatable.

Similarly, there are plenty of stories all over the web of people with all kinds of Macs, cables, drives, etc with a stable drive before upgrading macOS to Big Sur or newer, ramming into this problem... and then- because they need a stable drive more than bells & whistles of new macOS- DOWNGRADING macOS again and finding the same drive, through the same cable attached to the same Mac, etc, etc becomes perfectly stable again. What changed?

I'm towards certain the only actual fix is Apple getting in there and making the U in USB actually mean what the word means. In other words, we need some "just works" debugging of macOS. I suspect one day they will get around to debugging this code and then magically all these enclosures will become stable again. However, it's been now 5 generations of macOS so no one should hold their breath.

 

[EugW]

Are these USB 3 drives? I had this problem with several different USB 3 drives on my M1 Mac mini, but not with my two USB 4 / Thunderbolt drives on my M4 Mac mini.

It should be noted that macOS detects USB 4 / Thunderbolt drives differently than USB 3 drives, and reports them differently in System Report too. (If you plug a USB 4 drive into the M4 Mac mini’s front USB 3 port, it also gets detected and reported like other USB 3 drives.)

It would appear that USB 4 and Thunderbolt drives in macOS are implemented fundamentally differently than USB 3 drives. Unfortunately this causes issues with USB 3 drives, but on the flip side, I never get sleep disconnects with my USB 4 / Thunderbolt drives, and my Photos Library sync always works perfectly now with USB 4 / Thunderbolt.

BTW, I say “USB 4 / Thunderbolt” because if I connect it directly to the Mac mini’s rear port, it is detected as a USB 4 drive. However, if I connect it through a Thunderbolt 4 / USB 4 hub, it is detected as a Thunderbolt 3 drive. In terms of macOS functionality, USB 4 and Thunderbolt drives appear to work similarly, just with some differences in speed and power usage. In Thunderbolt 3 mode it runs a little slower than USB 4 mode, but Thunderbolt 3 mode also uses less power at idle, so I use Thunderbolt 3 mode. Photos sync works correctly in both USB 4 and Thunderbolt 3 modes. I only encountered the sleep and Photos sync issues with USB 3 drives. Both USB 4 and Thunderbolt NVMe SSDs also support TRIM in macOS, whereas USB 3 SSDs do not.

https://forums.macrumors.com/threads/thunderbolt-3-m-2-nvme-ssd-enclosures.2027925/page-79?post=33683940#post-33683940

I'm towards certain the only actual fix is Apple getting in there and making the U in USB actually mean what the word means. In other words, we need some "just works" debugging of macOS. I suspect one day they will get around to debugging this code and then magically all these enclosures will become stable again. However, it's been now 5 generations of macOS so no one should hold their breath.

See above. The "fix" came with USB 4... probably because USB 4 incorporates Thunderbolt protocols into it.