Unexpected realtime delay error message

how did you configure yor interface eth
#7i76e PL3 2019
auto eth0
iface eth0 inet static
address 192.168.1.4
netmask 255.255.255.0
network 192.168.1.0
broadcast 192.168.1.255
gateway 192.168.1.1
dns-nameservers 8.8.8.8 4.4.4.4
hardware-irq-coalesce-rx-usecs 0


do you got this inside your interfaces
sudo gedit /etc/network/interfaces

Thanks to everyone who helped especially PCW,

The new computer is working well. My old PC was probably missing the plastic fan duct and was overheating.

I've had the latest ISO running on a fanless J1900 PC for a couple of weeks.  Mesa 7i92 card (ethernet).      
I believe the shop air temp would be below 27C.  The J1900 box doesn't have a mount and has been placed at various angles.  I do recall latency test being ok, I will check it again.  This error didn't ever occur on the prior Dell PC running 2.6 RT Preempt, I'm assuming this is related to the new hardware.  It was an old Dell but low latency.  That Dell is still there.

This is a makerspace with many users and has been running tens of hours since upgrading, and this is something reported, I'm not aware of any unusual circumstances.  It threw this error once, and they had to restart their job:
"Unexpected realtime delay on task 0 with period 1000000
This message will only display once per session. Run the latency test and resolve before continuing. "

So, it's rare.  I don't know of a way to reproduce this, it just happened once and did not reoccur AFAIK.  It's only thrown it once in tens of hrs AFAIK, although I did hear someone else say "it just stopped once" for them and no more info about any error message.

Could this be J1900 overheating, or could it be this "IRQ coalescing" problem?   I'm leaning towards an overheating issue since it has only showed up rarely.

The only thing that comes to mind would be to TRY to overheat the PC by putting a blanket over it or something.  

Should I try to increase the SERVO_PERIOD=1000000?   I'm thinking not, as it operating ok for many hrs and it was fine with the old Dell.
 

It's an external wall wart, but I figured if anything went wrong with it, we'd see a reboot or freeze, not just an error message in LinuxCNC.

But I can imagine if it overheats, the BIOS might lower clocking to the system or some part of the system (like the ethernet), which could clearly create a latency problem.

However, if it were overheating, I'd also expect that once it did it, it would predictably repeat the problem in a few minutes after rebooting and restarting the job, if nothing else is changed.

The J1900 based mb I am using for mythTV struggled for years with random lock ups. It was mostly due to a kernel bug for that chipset. There is a range of Linux kernels that just won't work with them.

Ouch! Troubling! I spent a lot of time getting this right on this box.

Do you know of a better fanless mini PC for this?  Or any others I might want to avoid?

If I swap to another sort of PC, can I just swap the hard drive?  I did go through significant hassle recompiling the code for the VFD and deploying a bug fix.  Is it ok if it's another Celeron?

Generally yes. Possible exceptions if you used UEFI to install.