Mac Studio M2 Max external SSD boot slow with USB hub

Question

FidelS.80 Author

Level 1

5 points

Mac Studio M2 Max external SSD boot slow with USB hub

I have a Mac Studio M2 Max running macOS 27.0 “Golden Gate”, build 26A428.

The Mac boots from an external 4 TB Crucial P3 Plus:

SSD: Crucial CT4000P3SSD8, firmware P9CR30A
Current enclosure: OWC Express 1M2 USB4, 40 Gb/s
I replaced both the enclosure and the USB4 cable
APFS volume is approximately 92% full
SMART status: Verified

Booting from the internal Apple SSD is fast. Booting from the external SSD varies considerably:

With the USB hub and external backup drives disconnected: approximately 45–55 seconds
With the Time Machine drives connected: 1:45 to 2:40
Occasionally more than 3 minutes

The Time Machine drives are connected through a TP-Link UH700 powered USB hub, connected directly to the Mac:

WD My Passport 2628
Toshiba EXTERNAL_USB
Other Time Machine drives

The macOS logs show errors during launchd’s mount-phase-2:

My Passport 2628:
AppleUSBIORequest ... pipe stalled
MODE_SENSE_10 failed

EXTERNAL_USB:
AppleUSBXHCIRequest ... pipe stalled

Extreme SSD:
ClearNotReadyStatus in progress for 41 sec
InitializeDeviceSupport error

When the hub is powered on, boot can pause for more than a minute while mounting the external drives. With the hub powered off, boot returns to approximately 50 seconds.

I have also experienced:

kernelmanagerd stall: registry root held busy

as well as several shutdown stalls and kernel panics.

Has anyone experienced similar issues with Apple Silicon Macs, macOS 27, powered USB hubs, and Time Machine drives?

Do these errors point more strongly to:

a failing USB drive;
the USB hub or its power supply;
a USB controller compatibility issue in macOS 27;
or an APFS/Time Machine mounting problem?

What is the safest way to identify the problematic drive without risking the backups?

Mac Studio, macOS 27.0

Posted on Sep 21, 2026 12:55 AM

Reply

Answer 1

Top-ranking reply

PreCognition

Level 6

12,112 points

Sep 21, 2026 1:06 AM in response to FidelS.80

I’d look at the hub/backup-drive path first. The pipe stalled, MODE_SENSE_10 failed, and ClearNotReadyStatusmessages are USB/SCSI device-level errors, so they don’t point first to APFS or Time Machine. The biggest clue is that boot time returns to about 50 seconds with the hub powered off.

For the safest test, set Time Machine’s backup frequency to Manually, shut down, disconnect the hub, then test each backup drive one at a time connected directly to the Mac. Boot once with each drive attached. I wouldn’t run First Aid, erase, or reformat anything yet.

If one drive reproduces the delay when directly connected, I’d focus on that drive, its cable, or enclosure. If every drive works normally when connected directly but the problem returns through the UH700, the hub, its power supply, or a hub/drive compatibility issue becomes much more likely.

The external boot SSD being 92% full is worth addressing, but it doesn’t explain the minute-plus stalls that appear only when the backup drives are connected. SMART “Verified” also doesn’t rule out a USB transport or bridge problem.

I’d also see whether the shutdown stalls and kernel panics stop with the hub disconnected. That would help narrow it down further.

Reply

Answer 2

FidelS.80 Author

Level 1

5 points

Sep 23, 2026 3:59 AM in response to PreCognition

New data point that points away from the hub and directly at the external boot SSD itself.

This morning the Mac took ~3 min 55 s to reach the login window (kernel boot 08:16:44 → loginwindow 08:20:39). The enclosure is brand new (OWC Express 1M2 USB4, 40 Gb/s) and the USB4 cable was replaced. The powered hub was switched off and the other external SSD was disconnected — the Time Machine volumes only mounted after login (08:21:02+), i.e. they were not involved in the slow boot phase.

The APFS log shows the external boot volume itself (disk5, Crucial P3) spent 211.7 seconds trimming free space during mount:

08:20:31 kernel[0]: (apfs) spaceman_trim_free_blocks:4149: disk5 scan took 211.668888 s, trims took 207.880793 s
08:20:31 kernel[0]: (apfs) spaceman_trim_free_blocks:4153: disk5 99458661 blocks free in 1071243 extents, avg 92.84
08:20:31 kernel[0]: (apfs) spaceman_trim_free_blocks:4162: disk5 99458661 blocks trimmed in 1071243 extents (194 us/trim, 5153 trims/s)
08:20:31 kernel[0]: (apfs) spaceman_trim_free_blocks:4165: disk5 trim distribution 1:242347 2+:154039 4+:303365 16+:168319 64+:111744 256+:91429
08:20:31 kernel[0]: (apfs) spaceman_trim_free_blocks:4173: disk5 trims dropped: 3811971 blocks 738072 extents, avg 5.16
08:20:31 kernel[0]: (apfs) spaceman_fxc_print_stats:475: disk5 smfree 99458554/976750545 table 3853/3853 blocks 22412728 2325:5816:579746 22.53% range 118459:976565837 99.98% scans 1
08:20:31 kernel[0]: (apfs) spaceman_fxc_print_stats:494: disk5 scan_stats[2]: foundmax 579746 extents 1067383 blocks 77045971 long 2324 avg 72 77.46% range 165544:976585001 99.98%

So the external boot SSD performs a 1,071,243-extent TRIM at mount, at 194 µs per command = ~208 s — with no hub and no Time Machine drives connected.

For comparison, the internal Apple SSD trims the same way at ~1 µs/command and finishes in 0.05 s:

08:20:28 kernel[0]: (apfs) spaceman_trim_free_blocks:4162: disk3 41459748 blocks trimmed in 27544 extents (1 us/trim, 699317 trims/s)

The external volume is ~90% full, so its free space is fragmented into ~1.07 million tiny extents (avg ~93 blocks). The bottleneck looks like the external NVMe path (USB4 bridge + QLC DRAM-less drive) combined with the near-full, fragmented volume — not the hub or Time Machine.

Reply

Answer 3

den.thed

Level 10

98,292 points

Sep 21, 2026 7:34 AM in response to FidelS.80

In addition to temporally eliminating and testing without the Hub.

Make sure that the Startup Drive and Hub, are not on the same TB 4/USB 4 Bus.

For example: Plug the Startup Drive into port 1 or 2 and the Hub into port 3 or 4.

Also note: that if you need more rear ports, you can move one TM backup drive and/or the Hub with both to a front port.

Reply

Answer 4

den.thed

Level 10

98,292 points

Sep 23, 2026 7:44 AM in response to FidelS.80

FidelS.80 wrote:

The external volume is ~90% full, so its free space is fragmented into ~1.07 million tiny extents (avg ~93 blocks). The bottleneck looks like the external NVMe path (USB4 bridge + QLC DRAM-less drive) combined with the near-full, fragmented volume — not the hub or Time Machine.

1) An SSD needs at least 20-25% free space for optimal performance.

2) The P3 is Crucial's bottom of the line affordable M.2 SSD.

The way to prove that is to,

1) Free up 20-25% of space on the P3 SSD.

2) Move everything to larger Gen4 M.2 SSD.

FWIW I still recommend connecting an external startup disk, directly to the Mac instead of thru a Dock or Hub.

Reply

Answer 5

Sep 24, 2026 7:36 AM in response to FidelS.80

As den.thed has already said, running a drive (other than a Time Machine drive) more than about 80 percent full is a Bad Idea™.

TRIM is an instruction to the drive Controller that specific data blocks have been deleted, and the drive Controller is now able to add them back to its pool of available blocks.

It's interesting news to me that MacOS TRIM operations have been deferred until the next startup. But once you say it is working that way, it makes sense -- that drive's performance will not be delayed by "live" TRIM operations to a drive that may be needed for active reading or writing.

Deferring TRIM combined with a drive already very full, suggests that blocks are not being made available for re-use immediately after deletes, making the drive space issue even more dire.

Conversely, making more space available on that drive should be an easy way to improve performance.

I whole-heartedly support what den.thed suggests for a further experimentation.

Reply

Answer 6

den.thed

Level 10

98,292 points

Sep 24, 2026 7:50 AM in response to FidelS.80

Make more free space on that external startup disk.

Free up storage space on Mac - Apple Support

Make sure that external drive is set as the startup disk.

Change your Mac startup disk - Apple Support

Make sure that startup drive is not connected to the DFU port.

How to identify the DFU port on Mac - Apple Support

Make sure that the macOS version on the internal drive up to date.

Use more than one version of macOS on Mac - Apple Support

Contact Apple Support and escalate the problem up the latter.

Contact Apple Support - Apple Support

1-800-MY-APPLE (800-692-7753)

Reply

Answer 7

Sep 21, 2026 7:52 AM in response to FidelS.80

Time Machine can accept as many additional drives as you wish to attach. When there is more than one, each drive will hold a complete stand-alone backup set, starting from the day it was created. Every-other backup goes to every-other drive, round robin. Because each backup drive is independent, any drive can be removed at any time and all others retain their integrity.

den.thed suggestion to move the Time Machine drive to a front port if you need to is a good one. Time Machine does its backing up work at low priority in the background. Time Machine does not get any extra benefit from a drive or even a connection that is FAST.

Reply

Answer 8

FidelS.80 Author

Level 1

5 points

Sep 24, 2026 6:08 AM in response to den.thed

The NVMe drive is connected directly to the Mac; given current prices, upgrading the drive isn’t feasible. What I’d like to understand is that every time I boot up, there’s a different problem – sometimes it’s the TM drives, sometimes it’s TRIM, sometimes it’s the other external NVMe drive – so it’s not even clear which one I need to replace or where I need to change the settings.

I’ve always booted my Mac from external drives, ever since macOS 7 with SCSI drives; this random problem is driving me mad.

Reply