error finishing read

Yes, we are at 10.10.10.10.
Running patch   sudo ethtool -C eno1 rx-usecs 0
Intel device

Look at CPU history. It looks like this for weeks at a time except for failure, and then it goes spastic. Note failure time.

OK, that might be several things, so check the "processes" tab when that happens, see what is causing such CPU usage.
Another thing that can do such high usage is file system journaling, especially if the hard drive is on its way out.
Disable automatic updates, i do that always, also disable everything you do not need from the "startup applications" in "preferences" menu.
This can also happen at random if there is a broken PDF file somewhere on the drive, very hard to pinpoint, had this issue several times. Easiest way to find this out when opening file browser if the thumbnails take long to show or do not show at all for some file types.
And yes, these are all for Linux Mint, although other distros can have the same issues.

scotth wrote:

Running patch   sudo ethtool -C eno1 rx-usecs 0

 

This is only temporary and stops on  a reboot. Much safer to do it in the networking config file as per the hm2_eth docs.

I agree on the patch, but at this time I am the only one running the system and I know that I need to do it when I reboot.

Tommy,
My startup file is much shorter.

 

Did you disable AMT in the BIOS?

Yes, now all I can do is wait. It has only happened every week or two, so it may be a while. Thank you for the AMT info.

Watching

hm2_7i76e.0.packet-error-total
and
hm2_7i76e.0.read.tmax

Will give you more information sooner than waiting for a fatal
cluster of latency hits

linuxcnc is on ens34 ethernet card and does not run on isolated core.
How to change this ?
My ping is low but sometimes explodes to 1-30ms.
Motherboard is shuttle XG41 and has very limited access to bios options.
Mobo has 2 onboard realtek eths / driver is from realtek installed with dkms.