Only one restart is required, but you get a notification from SoftRAID, and also System Preferences.

With 6.3 the SoftRAID one is not modal, so you can restart from System Preferences.

Don't know if this is a bug or intended action... If the System Preferences>Security & Privacy>General Allow button is not checked and the Restart is selected from the popup (instead of the Restart after the Allow button is selected), the Restart has to be done again, since the Allow button remains unchecked. Is there a way around this dual restart? Has it always been thus? I usually go into Systems Preferences and do the restart from that notification. Maybe the bug should be titled, "dual notifications." Guess I'll report that separately. 

At this time, all we really have to offer is disabling sleep settings, including set display sleep to never.

If there are ejects, moving connections around, trying to find a combination that does not trigger disk ejects.

After that all one can do is replace various devices, or reduce what is connected. TB should be far better than this, it was designed to be.

@henry-in-florida 

One note that is easy to miss, there is another thunderbolt chip in each connector end, including each end of a cable. (no way to measure this with an app) As an example, fiber optic Thunderbolt cables are great for long connections, but heat causes the same eject issues and apparently it develops/gets worse with time, as after a year or two, they inevitably are no longer usable, as the disks eject so often. OWC used to sell them, but way too many returns.

I think you may not see this as much on non SoftRAID disks, as SoftRAID reports all events, it does not ignore anything. macOS may not report disk ejects if an app does not detect a change in the time a volume was off line.

Lets hope your configuration stays stable now!

@softraid

Very interesting. I don't have long cable runs, each is 0.5-1.0m long. If you have any other resolutions to this from your aspect, please let us know here. Understand that you're dealing with a macro issue affecting all kinds of use cases, not my specific single issue. 

I've been connecting/disconnecting every few days the T3 connection using the MBP off the desktop and doing manual battery optimization (about 30-40min/time) because the auto-optimization macOS Monterey software doesn't meet my battery readiness needs (another story, actually quite bad, it stinks and I've complained to no avail). There's plenty of in-usage testing on the cable ends at the MBP end, which are in great shape. The T3 connection is far superior to its predecessor, IMO.

One note that is easy to miss, there is another thunderbolt chip in each connector end, including each end of a cable. (no way to measure this with an app) As an example, fiber optic Thunderbolt cables are great for long connections, but heat causes the same eject issues and apparently it develops/gets worse with time, as after a year or two, they inevitably are no longer usable, as the disks eject so often. OWC used to sell them, but way too many returns.

I think you may not see this as much on non SoftRAID disks, as SoftRAID reports all events, it does not ignore anything. macOS may not report disk ejects if an app does not detect a change in the time a volume was off line.

Lets hope your configuration stays stable now!

@henry-in-florida 

the problem with this issue is it is a flaw in Thunderbolt. I think there is either interference, or perhaps a heat issue, where one of the Thunderbolt chips in the connectors crashes. Every connection point has a Thunderbolt computer chip on each end, which is why it is fast, but the connection is very susceptible to interference.

One would think after 7+ years of this happening on Thunderbolt buses, either Intel or Apple would figure out this is affecting the pro market (people with lots of drives/monitors, devices, etc.) and address it.

Or at least fix it in thunderbolt 3? Or 4?

 It shows how the little these companies consider what happens when users actually use their computers with lots of devices.

@softraid,

While we agree it is a problem and a likely hardware issue, as you also said, it's intermittent for me and heat is making a very good guess as to the cause. To add to that with some diagnostics and local info IF it is heat related or communication related (device, possibly cable), my problem has 'settled down' since last reports. 

I can monitor heat of every sensor independently with an app called Temperature Gauge (TG Pro app). The left and right Thunderbolt sensors run 56 and 67°C respectively. 67°C is about the the overall average temp of the machine with the clamshell closed 24x7, the normal condition of the MBP when on the desk and connected to the Thunderbolt devices. Why the left side of the entire machine runs cooler than the right side is a design mystery, however I tried to cheat the system in TG Pro and make the right side fan work at high speed with mixed results when the team won the right exceeds 65°. Not very effective (throttling issue with Intel machines?). 

Made the change to closed clamshell for a power savings when I got the new monitor hooked up. It works well, except I have no use of TouchID. BTW, the new keyboard with TouchID built-in Apple sells at a high price, only works with Apple Silicon equipped MBP (so not backwards compatible). Thanks, Apple! I can use it nicely on their 2020 MacMini M1, so it's compatible with that. Sweet keyboard, unfortunately not usable where needed most- on Intel MBP. I remember my logon though it's less than convenient. Just tryin' to get through until M1 or M2  is available for MBP. I don't like 1ˢᵗ GEN Mac hardware. Since we're ranting... 

The logs are clean since 14 APR (only two days, I know, but just sayin') events. I have not had any failure. No further reports there, nor have there been diagnostic reports of Thunderbolt 'glitching' or 'resets' you describe. Is that something that would be reported/logged without SoftRAID? Seems like any SSD would react in the same way, no? However, the other SSD's attached do not exhibit this issue. Why?

the problem with this issue is it is a flaw in Thunderbolt. I think there is either interference, or perhaps a heat issue, where one of the Thunderbolt chips in the connectors crashes. Every connection point has a Thunderbolt computer chip on each end, which is why it is fast, but the connection is very susceptible to interference.

One would think after 7+ years of this happening on Thunderbolt buses, either Intel or Apple would figure out this is affecting the pro market (people with lots of drives/monitors, devices, etc.) and address it.

Or at least fix it in thunderbolt 3? Or 4?

It shows how the little these companies consider what happens when users actually use their computers with lots of devices.

the fact this is happening with a new Studio Monitor means this issue has not been addressed yet by Apple. The cause, it appears, is one of the thunderbolt chips resets/crashes. this is not a problem for monitors, but Thunderbolt is designed to be "hot swappable", so disks power off instantly. that is why you get the disk eject messages.

@support

So far today, it has not failed,0 despite my taking the machine on and off battery power, disconnecting MBP from its normal Thunderbolt ports and reconnecting from a portable configuration several times. The devices came on line when the connection was plugged in and remained on line. 

I have two monitors connected to Thunderbolt on the machine, one on each of two Thunderbolt ports. The 5K new monitor is all by itself on one thunderbolt port, while the 4K monitor is attached to the port that loops through multiple devices (a hub, two OWC drive enclosures using all SoftRAID components and the monitor), as it has for a long while - stable as a rock until recently with MacOS 12.3.1, no trouble. 

I won't get into a fingerpointing exercise except to say that I filled in all the blanks for you on how its connected and the devices and what error flags are being raised. If you say on set of chips is failing then that's obviously hardware and it's under warranty. But I know that Apple is going to make me go through a bunch of hoops to prove and I cannot do soexcept to point at your drivers. So it won't get fixed. I'll keep reporting it to you until you come up with a way to show some other provable Apple hardware issue.

I was just commenting on how some users would see the monitor flash on occasion, showing that while the Monitor may also momentarily eject, it is barely noticeable, unlike disks, which is bad news.

the fact this is happening with a new Studio Monitor means this issue has not been addressed yet by Apple. The cause, it appears, is one of the thunderbolt chips resets/crashes. this is not a problem for monitors, but Thunderbolt is designed to be "hot swappable", so disks power off instantly. that is why you get the disk eject messages.