MP 7,1 Cursed Mac Pro 2019.... help?

[Dt990]

My Mac Pro 2019 keeps beachballing then logging out, thanks to WindowServer and Logd crashing repeatedly. Bluetoothd also is frequently involved. The crash log isn't helpful as it's curiously absent on causes. Generally the flowing info is missing or blank:

It's been happening a few months but wasn't frequent and got worse a few weeks ago.

The machine also boots very slowly, especially off NVMe.

Path: ???
Version: ???
Code Type: 00000000 (Native)
Parent Process: ??? [Unknown]
User ID:

Crashes can happen while the computer isn't in use (less common). It would reboot when Monterey sometimes. Ventura just logs me out. Some crashes cause the monitor to output garbled images.

Machine Specs:
Mac Pro 2019 8 Core / 64 GB of RAM / 6900 XT / 4x NVMe (between 3 PCIe hosts) / 2x SATA SSDs (internal) / Sonnet Allegro Pro USB Card / External USBc 4 drive Terramaster / External WD HDD / Focusrite Clarett / Samsung T7 that's shared between computers / Anker USB hub / Logitech mouse. Connected to two 4k Displays.
Running Ventura 13.4 was previously running 12.x when this problem started. I don't use much in the way of system utilities.


Here's what I've done, not in order:

• Removed old system junk (launch apps / kexts). No change.
• Removed all 3rd party Kexts (including Logitech drivers). No change.
• Tried disabling Little Snitch. No change.
• Moved my system from aging Samsung 970 Evo to Samsung 980 Pro. No change.
• Upgraded to Ventura. Stopped any reboots. Now only logs out.
• Removed the 6900 XT and put the factory MPX 580X back in. No change.
• Reset NVRAM... many times. No Change.
• Pulled the Ableconn 2 SSD PCIe card. No change.\
• removed the rarely used Bootcamp SSD. No change.
• Unplugged SATA SSDs. No change.
• Downloaded and ran Malwarebytes anti-malware. Nothing found.
• Tried Disk utility on my drives. Nothing found.
• Pulled the two 16 GB OWC DIMMs. No change.
• Unplugged my Clarett. No change.
• Ran color profile verifier (this gave me grief years ago when a color profile got corrupted). None found. Deleted a bunch of created ones for good measure. No Change.
• Swapped two of the factory DIMMs out and put the OWC RAM back in place of the two of the DIMMs (possibly one or more DIMMs are bad). No change, but I haven't tried swapping the other two factory DIMMs. \
• Tried diagnostic mode (booting holding d key). I did get an error code that said, "Possible problem with fan". Tech Tool Pro said fans were fine. Temps are pretty cool. Generally 50c.
• Tried booting on the internal SSD and found I still had the problem. This is the most concerning, as the internal SSD has been virtually untouched. I upgraded it some time ago from Big Sur to Monterey, and I have Tech Tool Pro on it and Firefox. That's about it. I still managed to WindowServer crash.

Future Actions:

I've ordered 4x 32 GB Crucial RDIMMs off eBay, so I can at least test it with a totally fresh batch of DIMMs. Those won't be here to until next week. I'm going to try removing everything beyond the factory USB/Thunderbolt card. I still somewhat suspect the RAM as I had ram go bad in my Mac Pro 3,1, but it was more catastrophic, causing random kernel panics.

I haven't tried a restore, as my internal SSD is essentially factory.

I'm trying to avoid the Apple store but I may have to give in. Apple care is expired.

 

[Matty_TypeR]

I would remove all and go back to the base machine config. fresh install of the latest ventura 13.4.1 as a starting point. use the machine and see what happen's. if its ram based fault it will show as base config. If all is ok add the next part one at a time and use. It might even be a fresh install will cure your issues but return to base config and fresh install to start with.

 

[ZombiePhysicist]

Agree with the above but would go further. Not just fresh install, but fresh empty user account. You may have some cruft in your user account causing the issue.

 

[avro707]

I think a complete fresh install with nothing else on it at a base config is worth trying. See how it runs for a week.

I tried to search some of the errors but they don’t seem related.

 

[Dt990]

Well I reinstalled macOS from recovery mode and restored via Time Machine on Sunday, even during the Time Machine (Data but did not copy system settings), there was a series of mds_stores that happened over night and while trying to update to Ventura the next day, I got a series of PerfPowerServices errors, and finally just got a logd. It seems stupid and laggy just as before.

The PerfPowerServices crashes were:
Exception Type: EXC_CRASH (SIGABRT)
Exception Codes: 0x0000000000000000, 0x0000000000000000
Exception Note: EXC_CORPSE_N.OTIFY

and the mds_stores

Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00007fb900000110
Exception Codes: 0x0000000000000001, 0x00007fb900000110
Exception Note: EXC_CORPSE_NOTIFY

Both of which being part of launchd so I trashed all the /Library and ~/Library launch agents and launch items,

The logd was the typical lack of info. The only things I had in my computer were the 580x, and a new 980 Pro in a PCIe host, and 4 DIMMs (rotated out the two factory DIMMs). RAM doesn't come until end of week if I'm lucky.

Something well and ****ed here, I only attached my USB HDD to restore.



EXC_BAD_ACCESS (SIGSEGV) is segmentation fault and EXC_CRASH (SIGABRT) which is a memory run time error. I'm not a memory managed sort of dev but aware of these errors. So it's nice that this go around I got something here and a ****ing process to at least attack. Still, not great that even RESTORING from Time Machine and my native SSD I still random issues.

I'm crossing my fingers that I have like bad RAM, like a DIMM or four as with post SIP and especially now that we're at the level of metadata integrity protection, and the OS being distroed as a damned image read only /System, that it cannot be something wrong with macOS itself due to a few bad bytes on the system.

 

[tsialex]

Did you run Apple Diagnostics?

Use Apple Diagnostics to test your Mac - Apple Support

Apple Diagnostics, formerly known as Apple Hardware Test, can check your Mac for hardware issues.

support.apple.com

 

[Dt990]

I did awhile ago and i received an Fan error notification but nothing that seemed catastrophic. TechToolPro has a fan test and they spin and the heat seems all within normal temps, often crashes happening when most components are 50C or lower.

 

[DrEGPU]

Could be a PSU failing. Random problems like that usually are either RAM or PSU. It could certainly be something else of course

 

[avro707]

I did awhile ago and i received an Fan error notification but nothing that seemed catastrophic. TechToolPro has a fan test and they spin and the heat seems all within normal temps, often crashes happening when most components are 50C or lower.

Did the other ram arrive? Did it help or no difference?

 

[Dt990]

Did the other ram arrive? Did it help or no difference?

Got 128 GB of Crucial in the system.

I'm day #2 without any system level crashes (the audio crashes are fine as it's FCPX validating plugins). Since the Time Machine restore I haven't seen any issues. Today I dropped in the 4x NVMe + Sonnet Allegro Pro + 6900 XT + plugged in all the USB devices.

For posterity if someone else finds this thread and experiencing repeated logd / windowserver crashes and bluetoothd.

We shall see how it performs. I'm guessing this was bad RAM or some errant service that I nuked when I went on a /Library rampage, deleting all LaunchAgents / LaunchItems and clearing out application support for anything super old. I also reinstalled most of my audio plugins (hence the crashes as FCPX tried to figure out which worked). Unlikely the audio was to blame but what a pain the goat ass. That was more brought on as I needed to re-authorize some of them after Time Machine restoring thanks to aggressive copy protection.

If it continues to run smoothly, I may chance dropping the two OWC DIMMs. It'd be nice to have a bit more RAM but 128 GB has served me well. Hopefully this stability continues.

/edit: The funny thing is this terrible behavior correlated with the time when I tested an exploit in Apple's system firmware challenge that a reddit user. It's always possible that the Find My Mac was forever ****ed after that causing logd and windowserver to freak out.

 

[Dt990]

.... and maybe not fixed. ****ing hell.

 

[cobra521]

Dt,

Probably you've already done this, so forgive me if you have. I searched "userspace watchdog timeout windowserver ventura" using DuckDuckGo and got numerous hits.

Sadly, it also looks like none of them have solved this problem. Their suspected causes seem to be all over the map.

Having said that, maybe you can find one that has a suggestion that could help you if you go through them. Not much help, sorry,

Tom

 

[Dt990]

Dt,

Probably you've already done this, so forgive me if you have. I searched "userspace watchdog timeout windowserver ventura" using DuckDuckGo and got numerous hits.

Sadly, it also looks like none of them have solved this problem. Their suspected causes seem to be all over the map.

Having said that, maybe you can find one that has a suggestion that could help you if you go through them. Not much help, sorry,

Tom

Yeah, I just had a kernel panic movements ago. I did see the suggestion of hunting down the windows server plists which isn't a bad move, also dumped some plists that go back as far as 2005 just for the sake of house cleaning. A lot of the suggestions with system prefs I already set as I don't like screens as separate spaces and dislike mission control immensely. The only thing I don't have set never turning off my displays.

Outside of some dev tools (the usual smatter of CLI apps via homebrew, docker and web dev stuff) and audio stuff, I'm not exactly exotic with my macOS use.

 

[Dt990]

Decided I'll try this as a live blog since I'm so deep in the woods here. Maybe someone else can see the forest or that I'll solve it and someone else will benefit:.

Finally found my first repeatable crash:

Onyx crashed running the default Maintenance. I booted into safe mode and ran Onyx (sans APFS snapshots) and it was able to complete. Once it completed and rebooted, my Mac Pro went to a kernel panic. It displayed the message twice, once after the chime and once after the boot.

Prior I'd have to use my computer aimlessly for hours until it'd crash, at least now I have a repeatable way to test against. Reboots are incredibly slow. Between clicking reboot in Onyx, took 11.5 minutes to get to the login screen.

For the first time a KP wasn't a logd or windowserver, this was a /usr/lib/libSystem.B.dylib

panic(cpu 8 caller 0xffffff8009926f6e): initproc failed to start -- exit reason namespace 6 subcode 0x1 description: Library not loaded: /usr/lib/libSystem.B.dylib Referenced from: <EF26A3BB-B918-30CF-8F28-0AC45F932E67> /sbin/launchd Reason: tried: ‘/usr/lib/libSystem.B.dylib’ (no such file), ‘/System/Volumes/Preboot/Cryptexes/OS/usr/lib/libSystem.B.dylib’ (no such file), ‘/usr/lib/libSystem.B.dylib’ (no such file, no dyld cache)

Fascinating this was a dylib error of being not found.

Perhaps clearing out my logs might help the logd crashes. I'll have to try running Onyx with more extreme cleaning. I also will have a video on this.

 

[Dt990]

Latest update:

I'm back to seeing the same crashes after running Onyx. I had a glimmer of hope trashing the WindowServer prefs and dumping all the logs might fixed it. I really hoped swapping the GPU and RAM would have fixed it.

It's almost always the same, timing out and then crashing. I think I might have to go to Apple with this and am dreading the fact I don't have Apple Care. All I can figure now is there's a faulty PSU or logic board or such.

 

[Shawn Moosa]

Try remove every component except the CPU and do a deep cleaning.
Saved my machine from random hangs and extreme slow downs.

 

[Matty_TypeR]

You should try a fresh install with out using time machine. No point backing up from time machine as it may well contain the very thing that gives you KP's


Fresh install with standard factory machine layout. then add one part at a time and test.

 

[Dt990]

I'll likely to a full restore to the factory SSD but considering it managed to crash even when booting from it and it's virtually untouched, I'm going with something not software related.

 

[tsialex]

I did awhile ago and i received an Fan error notification but nothing that seemed catastrophic. TechToolPro has a fan test and they spin and the heat seems all within normal temps, often crashes happening when most components are 50C or lower.

Install back the Radeon Pro 580X, do a factory restore then and check Apple Diagnostics again, PPF004 is likely an error caused by third party GPUs.

If no error is found, test with this install - without Time Machine - for some days.

 

[nisse32]

I had the same problem (reboots, crashes, slow downs etc). I changed RAM, processor, graphics cards etc, no success. I also tried three different 980Pro 2TB boot nvme SSDs. The same problems continued.

For some reason it seems that I had to boot from the original Apple SSD, the 980 Pro would not work correctly. I bought another 7,1 with 2TB SSD (cheaper than trying to find Apple original separate upgrades) and this computer is working OK for the last two weeks. I will not spend more time to try to find a working PCIE boot SSD solution...

(Also my original 7,1 with 1 TB Apple original SSD boot drive worked OK but I really needed 2TB...)

 

[Matty_TypeR]

I had the same problem (reboots, crashes, slow downs etc). I changed RAM, processor, graphics cards etc, no success. I also tried three different 980Pro 2TB boot nvme SSDs. The same problems continued.

For some reason it seems that I had to boot from the original Apple SSD, the 980 Pro would not work correctly. I bought another 7,1 with 2TB SSD (cheaper than trying to find Apple original separate upgrades) and this computer is working OK for the last two weeks. I will not spend more time to try to find a working PCIE boot SSD solution...

(Also my original 7,1 with 1 TB Apple original SSD boot drive worked OK but I really needed 2TB...)

I have 3 samsung 980pro M2 drives used as boot drives, 2 mounted on a dedicated PCIe M2 card, and one stand alone PCIe cheap bifurcation card. I have no issue's booting from them all, Monterey on 1 drive, win 11 pro on another. and a back up of ventura on another. I dont have beach ball's or crashes on any of them.

The back up OSX version's do load slightly slower than the Apple SSD but win 11 pro loads faster than them all.

So i dont think the issue is a drive issue unless that drive is faulty some how.

 

[Lochnivar]

Watching this thread with interest. This has just happened to my A2141 i9 MBP 2019 16" and it's driving me crazy. Followed most of the steps you have, apart from going back to factory settings. I'm starting to think it's a hardware failure, and judging by the issues I've read, I'm not the only one. I will have to get hold of Apple too, I'm out of warranty. £2.5k for three years worth of a flagship mbp. Consumer law protects us in the UK and I will be citing it and/or taking Apple to small claims court if they put up any barriers.

 

[Rimmsi]

I have 3 samsung 980pro M2 drives used as boot drives, 2 mounted on a dedicated PCIe M2 card, and one stand alone PCIe cheap bifurcation card. I have no issue's booting from them all, Monterey on 1 drive, win 11 pro on another. and a back up of ventura on another. I dont have beach ball's or crashes on any of them.

The back up OSX version's do load slightly slower than the Apple SSD but win 11 pro loads faster than them all.

So i dont think the issue is a drive issue unless that drive is faulty some how.

Since I don't know the diagnostic tools for MacOS, try to install Windows 10 on the Apple SSD using BootCamp on the largest partition possible, then download HDTune Pro. In the HDTune settings, change the size of the tested blocks to 8 MB. Then do the Error scan, Benchmark and Health Card. Please post the results in this thread.

EDIT: Sorry, the quote should have been to the Dt990 user.

 

[Matty_TypeR]

Since I don't know the diagnostic tools for MacOS, try to install Windows 10 on the Apple SSD using BootCamp on the largest partition possible, then download HDTune Pro. In the HDTune settings, change the size of the tested blocks to 8 MB. Then do the Error scan, Benchmark and Health Card. Please post the results in this thread.

I don't have any issue's with any of the samsung 980pro M2 drives i use, no error's no crashes, all run perfect.

I think you might have quoted the wrong post here. I have no need for any tool's or diagnostic's.

 

[ZombiePhysicist]

These mystery weird situations are fascinating to watch. 🍿

I hope you track down the culprit!