Since a few weeks (about since 6.6.0) I'm repeatedly loosing my network connection after a uptime from one to a few days. From that moment on I get "r8169 0000:0f:00.0 eth0: link up" spamed in the syslog.
I can manually put eth0 down and up with no success, also plugin in to another switch doesn't help. I can shut down eth0 and start using eth1, I currently have no bridging or bonding active. (A side note: isn't it strange that I can still connect to the IP of eth0 even it is down?).
The same thing happens when I switch eth0 with eth1.

Hardware:
Mainboard Z68 Professional Gen3, I2500k with 2 onboard NICs (r8169),
16 GB RAM
Unraid 6.6.2

 

dednas-diagnostics-20181019-1624.zip

There's a known issue with Realtek NICs and latest kernels, though it's supposed to be better on v6.6.2, for the future you might consider getting an addon Intel NIC.

Thank you for the answer. I could not find that myself. At least glad to hear it is already known.
Since I never had any other problems with these NICs I'll just wait for a fix or maybe downgrade in the meantime.

19 minutes ago, dedi said:

Since I never had any other problems with these NICs

There have been a lot in the past, hence why I suggested getting an Intel, since this is not the first time, but hopefully it will be fixed soon.

Thanks, I'll keep that in mind

I'm still curious, why can I connect to an IP that isn't up?
My config:
eth0: 192.168.0.20, after the problem I shut it down with (ifconfig eth0 down). eth1 takes over and network starts working again.
eth1: 192.168.0.21
On other devices on the network, I still can connect to 192.168.0.20