Mac mini server periodically dumps all connected file-sharing users

This has been an ongoing problem since I set up this server about a year ago. It comes and goes, sometimes it happens multiple times a day, sometimes it goes stretches where it has no problems at all. Here is the setup:

A new 2014 Mac mini with 8 GB of RAM and a 250 GB SSD drive. The SSD is the boot drive and the storage drive where the shared folders live. There’s also a 2 TB HDD in the hard drive bay that stores Time Machine backups for the other machines on the network.

There are at most 3 users in the office, connected to both of the two shared folders, plus all these Macs use the server’s 2 TB drive for time machine backups.

There’s another two people that connect to the server’s VPN and access it that way at night sometimes.

The server runs file sharing for these two folders and the backup folder. It has its built in VPN server activated with appropriate port mapping so you can connect through their static IP and get right to the server’s VPN. It also runs FileMaker server, and apache/mysql. But they only run a very simple and very low traffic website on the machine.

All of these machines are connected via ethernet to a gigabit switch. It’s a pretty simple setup that should run very well, especially since the mini is running on SSD.

The Problem:

So what happens is, without any warning, the clients will all at the same time, get the beachball. They constantly work with files on the server, opening them directly. So what seems to be happening is that the server is loosing its network connection. Unfortunately I’m never on location when this happens so it’s hard for me to diagnose it. I thought it might be a bad ethernet port or switch. So I switched the server over to Wifi as a test. The problem seems to still happen. Everything else seems to be fine when it happens. Also this machine was on 10.13 High Sierra when I first set it up. I have since upgraded it to 10.14 Mojave and it has the exact same issues. They have to reboot the server to get file sharing back up and working, they are unable to reconnect to a ‘dumped’ shared folder. The server itself is not beach-balling or otherwise locked up. They are able to reboot it easily.

It almost feels like something is trigging a firewall on the server that is causing it to suddenly block incoming connections. The server is behind NAT so it’s firewall is (should be) turned off entirely. This is a very strange problem.

What I’m not 100% sure about is, if the server can still access the internet during one of these episodes, and if the connected users can still access FileMaker Server during one of these episodes. I think they can but I’m not certain because I only hear second hand reports from non-techy people.

Switches can always go bad, and this office has had issues with questionable hardwiring in the past. But if the problem was the switch or the ethernet runs, moving the server to wifi like I did should mean that clients would have problems SEPARATELY, and not at the exact same time.

I also make a script that runs every 90 seconds and logs the current local IP address, and the result of one ping from the server to it’s router, and from the server to my server (over the internet). I was hoping to get some clear results from that, but the results seem to be fairly random when an “episode” is happening. It never seems to drop its IP (via DHCP). But sometimes it can’t ping it’s router, but it can ping a server across the internet. Sometimes the opposite is true. Sometimes both pings fail.

I don’t think there is a hardware problem in a Mac that can cause such a specific issue, over both wifi and ethernet, while otherwise running perfectly 24 hours a day. It seems like it has to be either some kinda of weird hardware glitch, or some kind of infrastructure problem at the office that is manifesting itself in unusual ways.

Also note that Server.app is not running on this machine. Why bother since they removed all services from it.

If anyone has any thoughts about this, please share!