Obnoxious intermittent SMB issues on mojave Mac Mini (2018) server

[veranmed]

Ever since we installed this new Mac Mini (2018) we keep having intermittent SMB issues.

Setup: Mac Mini (2018)
Currently on 10.14.2 and haven't updated to .3 yet because I was hoping I could get some reliability before adding another variable
2TB internal flash storage (yes... we paid for the bigger storage) - This drive has the majority of our shared server files
32TB G Tech Studio XL (RAID 0) formatted APFS drive that houses 2 larger shared folders (I know APFS is a faux paux for external drives but the server refused to recognize some of our permissions when I still had it formatted Mac OS Extended and APFS seems to work)
16TB G Tech Studio (RAID 0) formatted Mac OS Extended running time machine on the entire Mac Mini storage and 32TB G Tech Storage
DNS enabler running DNS services
Mac OS Server App

Issue #1: Completely randomly SMB access seems to stop working. Remoting into the server via ARD reveals that all seems to be running fine but turning off OD and files haring and turning back on does not solve problem. Restarting the Mac Mini seems to be the only fix. I have had it running successfully for 10 days, for 5 days, sometimes less than 24 hours though before the issue comes back. We have Mac and PC users connecting and when things are working, they work flawlessly and quite speedy actually for both users.

Issue #2: When restarting the Mac Mini, it comes back up immediately of course but users are unable to connect in to the SMB shares for an unspecified period of time, somewhere between 0 minutes and about 7 minutes. During this time the connect to server shakes as though a bad password has been entered but then eventually starts working. Also notable during this time is that these same users can authenticate via our SonicWALL VPN during this time which is tied into our OD ... leading me to believe OD is not the issue. In the logs, SMB seems to start working after the following line appears in the logs: DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock (-30994)

See text from attached log below from before I restarted the server for it's SMB issue this morning at ~09:00. I removed the Google Keystone agent from starting after this log since it's polluting the log at the very least and we really don't need to make sure Chrome is auto updated on this machine. https://drive.google.com/file/d/1udoiCKWM8f4HmhFUUxnNBP0Kn2ZGAfHh/view?usp=sharing

See text from attached log below from after I restarted the server for that SMB issue (look at this one to see why SMB is not allowing access).
https://drive.google.com/file/d/1w-k4OBcO7q7A6kxDH4Av26Z4GsDz3qP9/view?usp=sharing

Thank you all in advance for any potential solution you may have to offer!


Let me know if you need any more info!

[veranmed]

Server has been more steady lately, with SMB only crashing occasionally. When this happens, AFP is of course still working fine. I have figured out the following procedure to reliably restore things to working order - restarting and waiting unknown amounts of time for things to start up again was not cutting it haha:
1. Restart the server
2. Turn off OD
3. Turn off SMB and AFP by unchecking the boxes under file sharing in the sharing pane of system preferences
4. Restart
5. Turn on OD
6. Turn on SMB and AFP reversing procedure above

After 2 crashes within 2 hours yesterday - 2:00PM CST and 4:30PM CST, I decided to go ahead and update from 10.14.3 to 10.14.4 last night, but have a feeling that won't do anything. See the following for the logs from yesterday: https://drive.google.com/file/d/1QaS8ejvFXcAY1bK5WnT3IIP3CYqKVYKr/view?usp=sharing
https://drive.google.com/file/d/1JJjiCOullKjIgO1bUAz_PG_AslXNZZ5A/view?usp=sharing

I will also add that I switched all of our shares to the 32TB External Drive which is now formatted Mac OS Extended. There is no longer anything shared from the internal server flash and nothing shared from any APFS drives.

Can anyone advise?

[CAL4D]

Hi Veranmed,

I'm experiencing a similar issues - SMB service fails / access stops on a frequent (daily) basis. I found CLI operations which seem to recover the situation without having to restart the system but this doesn't get to the underlying cause and I'm just wondering if you made any further discoveries?

FYI. Here's the CLI operations I use:

sudo launchctl load -w /System/Library/LaunchDaemons/com.apple.smbd.plist

sudo launchctl unload -w /System/Library/LaunchDaemons/com.apple.smbd.plist