I have this problem in three different servers at this time (Mac mini with 10.14.5). In another one with 10.14.4 no problems at this time.
Only happen with smb connections, not with afp. Yes, if your disk is changed to AFPS you can't do nothing about it (other people has talked about it in this discussion)
In my opinion we can't do nothing until Apple launch a new update solving this problem (You can downgrade, but that can not be possible for data loos).
Now I'm using an "emergency solution" until real solution from Apple appear. A shell script restarting the smb service and launchd file to start this script and execute every x hours.
Yes, I know it's not a best option but it's more than nothing.
Script (I saved this file with "smb-restart.sh" name:
#!/bin/bash
#
# Author: Angel Bonet
# 10/06/2019
#
# Stop service
echo yourAdminPassword | sudo -S launchctl unload -w /System/Library/LaunchDaemons/com.apple.smbd.plist
wait $!
# Start again service
echo yourAdminPassword | sudo -S launchctl load -w /System/Library/LaunchDaemons/com.apple.smbd.plist
wait $!
# End process, create log
echo `date` " " `whoami` " smb service restarted."
(yourAdminPassword must be changed by your admin computer account password to use sudo privileges in script execution)
Launchd file:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<!--
You should customize the following placeholders:
- smb.restartDaily.service : Service name
- SCRIPT_PATH : path where script file is located.
- LOG_PATH : path of log files
-->
<key>Label</key>
<string>smb.restartDaily.service</string>
<key>ProgramArguments</key>
<array>
<string>/Users/admin/Documents/scriptsShell/smb_restart_script/smb-restart.sh</string>
</array>
<!--
set RunAtLoad false if you don't want that script starts
automatically every time your mac os x boots
-->
<key>RunAtLoad</key>
<true/>
.
.
.
you can see the total files in github (https://github.com/angelbonet/smb-service-restart)
As you can see in calendar interval execution, every day at 05:30AM this script execute and service is restarted.
At this point, the users are reporting more discconecting issues at different hours in same day and I'm going to change this execution to loop of maybe three or four hours (experimenting until I catch it... you know).
When Apple release an update whith solution (Apple, please do it now, please... ) I'm going to deactivate this launchd.
Good luck...