Slow Boot To Desktop

I have a Mac Studio system that starting with the 26.x updates is experiencing very slow boot to my desktop. This system is usually turned off at night and rebooted the next day. I get the password prompt quickly, I enter my password, and the progress bar freezes - with no movement - sometimes for three or four minutes before the desktop appears. Then, the second thing happens: the efficiency cores go to 100% and I can’t do anything until this calms down - it usually takes three or four minutes. And I mean I can’t do anything: copying an empty folder from Pictures to the desktop gives me the progress window “Copying 3 items to “Desktop” (6KB) - and that sits frozen until the efficiency cores free up. I have tried to get a snapshot of what is running - and the culprit seems to be “siriactionsd” - but I also see kernel_task, mobileassetd, identityservicesd, etc.

My guess is that Apple has been sneaking in some OS 27 features to test. I know Siri needs to build a search database but it seems dumb to hit the system so hard on the initial desktop.

The obvious solution is to just sleep the system when I am through - waking from sleep is fast the next day - so that is what I am going to do. But it worries me some that this is something we are going to have to deal with in 27. I haven’t seen any discussions about how much prep work is going to be needed to make Siri (and the AI features) work. I do have a lot of data on this system (4TB internal, 14 TB of external SSDs (including 4TB for Time Machine). The boot delays seem to be worse if I haven’t used the system for several days - indicating to me that the indexing and other prep work is more extensive after several days. I have a MBA that isn’t used a lot and I am going to put the Public betas on that system to see what happens.

I’d be interested in any comments. Maybe I am missing something,

David

I’m sorry, I don’t have any productive comments. I just can’t stop singing your subject line to the tine of “Slow Boat to China”.

Maybe sharing that will get it out of my head.

Yes, basically I just hope it will correct itself in the future. (The downside to sleeping the system is that on wake I often set the error “you didn’t eject the drive before removing it” when in fact it is still connected. It just disconnected on its own. I have to physically disconnect it and plug it back in to get it to mount.

My other (long time) complaint is with how Mac OS handles I/O for big file transfers, It I want to duplicate a SSD and try to just copy it (3TB) it usually fails (after running for several hours). I know now that the best way to copy is to just run ChronoSync “backup” to the empty drive. The I/O is slower (more overhead) but it works. A developer friend says that Apple hasn’t paid much attention to this for years,

David wrote:
“It I want to duplicate a SSD and try to just copy it (3TB) it usually fails (after running for several hours).”

I could be wrong and welcome correction, but…
… Seems to me that during a “large copy” if the finder encounters a corrupt file, something that gives it trouble while trying to copy, the finder will just abort the ENTIRE operation. The whole job. Everything.

I would suggest CarbonCopyCloner for big jobs like that. When CCC encounters a corrupted or damaged file, it won’t quit. Instead, it will “continue around” the bad file, and save the filename into a log you can examine after the job is done. Thus, if you try to copy 3tb worth of files and only a few are bad, the copy/clone will go through anyway.

I don’t use ChronoSync, but it seems to work for you. I’m going to GUESS that it does pretty much the same thing with bad files as does CCC.

These are not corrupt files - the I/O just gives up. I do use CCC as a backup - but ChronSynch lets me synchronize files across volumes, It alerts me if if finds a corrupt file, When I am copying large folders I can make it work by copying sub-folders one at a time - but who wants to do that? I have had it fail, and I just stop, and then look at the two volumes and continue with copying the folders missing on the target volume. It’s just one of the things I have learning how to avoid.

David

A few things:

  • What if you log into a clean test account rather than your main one?
  • Have you tried swapping cables on your external SSDs so you can avoid shutting down the Mac, which shouldn’t be necessary? That seems like the primary problem to solve.
  • You could try working with a chatbot to figure out how to look in the logs to see what’s happening.

That has happened before. If I recall correctly, minor upgrades to OS X 10.12 Sierra (or maybe 10.11 El Capitan?) took a really long time. What was going on was that every upgrade secretly did a trial conversion of the startup drive from HFS+ to APFS.

You may want to try starting it up in verbose mode, so you can see the log messages while it is booting. On an Intel Mac that would just be holding Command-V keys, but that doesn’t work on Apple Silicon. Google says that you need to use a terminal command to switch it on: sudo nvram boot-args="-v".

I just bought a Studio Display XDR and connected it to the Mac Studio. Now, when I boot the system I enter my password and the screen goes completely blank. I just wait and I eventually (after a couple of minutes) I see my desktop. That makes me wonder if the monitor is an issue given how these Studio Displays have their own firmware (which I had to update when I first connected the display).

I have been sleeping my system to avoid the delay but then I have one external SSD (Samsung 8TB) that often (but not always) disconnects while the system is sleeping. Then I either disconnect the cable and reconnect it or just reboot. Checking the cables is an interesting possibility but I have been using the cable that came with that drive. I could look at the logs but haven’t because it is too much work.

A long shot is you could try re-selecting the Startup drive in System Settings.

It used to be that what could happen is no volume is explicitly selected, so instead when you startup it goes through a process to discover the possible drives and then pick one to use by default. That could cause a slow startup.

I say this is a long shot because I don’t know if this is still a thing with Apple Silicon, nor if the fact that you get to the password prompt means it has already picked the drive.

In terminal on my mac:

ps -ef|grep siriactionsd

  502   730     1   0  9:33AM ??         0:15.25 /System/Library/PrivateFrameworks/VoiceShortcuts.framework/Versions/A/Support/siriactionsd

Seems this has something to do with VoiceShortcuts.framework.

Adam suggested trying a new cable. I bought a TB5 certified cable and connected that. No dropped volume while sleeping for several days - so I thought that was it. But yesterday when waking from sleep I got an error message saying the system couldn’t access that volume and did I want to initialize it. I said “No” (of course) and rebooted the system, and everything was working (though I still had to wait a couple of minutes with the black screen before the desktop showed up). So that remains a mystery.

Maybe it is the size - it is a 8TB SSD with a little more than 6 TB of data.

I am going to do some more research to try to find out what is going on.

David

Have you run Disk Utility on that volume? It feels a little wonky to me.

Yes, I ran First Aid on the volume and it passed with no errors. I haven’t seen any ejected volumes or errors and I have been sleeping the system for several days. The error isn’t consistent. I also have DriveDx running on the system and it says the volume is fine.

I have 4TB SSDs that I used to use but ran out of TB/USB ports on the Studio so I consolidated the data on the 8TB SSD. I can use a hub (and have one on the system now). My Mini (that I use as a file server) has two hubs! Lots of wires on that system. Too much data that I am trying to keep online I guess.

David

I suppose there is a question of how much data needs to be attached at all times. :-)

Yea, I know! I have a lot of photos - and I shoot raw+jpeg with cameras that create really big raw files (e.g. Sony A7R cameras and Fuji Medium Format camera). The the images that are still “active” in some fashion are in iCloud (about 1.2 TB) and I have 4 TB of internal SSDs for both my Mac Studio and my Mac MBP16 - so those images are “local” too. But I have the big external SSDs so why not? If I have any complaint about Mac OS it is that it doesn’t seem to handle the really big data folders very well. Even just regular copies in the Finder of one SSD to another can fail - and often do. I generally use ChronoSync for these copies - a bit slower but they work!

David

I got another “Drive Unreadable” message after waking from sleep for the 8TB SSD again. (The drive is fine.) I have another hypothesis that I am going to check. This drive is connected directly to one of the four TB/USB ports on the back of the Mac Studio (M4). One possibility is that the Mac is not supplying enough power for the connection (or it is very marginal). I have moved the drive to a powered dock (which I know provides more power) and see if that helps.

I know power is an issue with Macs (not just the Mac Studio) and if you use all the ports for devices that need power, then you can have problems. (On my M1 Mac Studio I used to get error messages saying too may devices were connected.). This power issue also occurs for one of my cameras - a Fuji GFX100SII Medium Format camera. The Mac doesn’t recognize the camera if the USB cable is connect directly to the Mac Studio - but it works fine if the connection is to the powered dock. (This camera is unusual - I have other cameras that don’t have this power requirement.)

So, I will see if this fixes the issue.