Unraid 6.11.5 - Networking randomly dies

[jbat66]

This is a wierd one. I was away from the server for several weeks and when I came back the server was running, but no gui, or any network access. I rebooted from console, and all was good for about a day and then no network. the server was fine, just no network. I was running 6.x something and figured I would upgrade to latest version so upgraded to 6.11.5. Same thing, the server would run fine for a day or so and then no network. I then figured I would replace my NIC, but no more slots, so I plugged in a USB3 Ethernet dongle, and moved ETH0 over to the dongle. Everything stayed running for over a week. I then thought I would try the internal NIC again and set it up as ETH0, but left the USB3 dongle plugged in, and added another IP so I can get to the GUI if the NIC/networking quit. Yep that night, networking quit, for both. I had to again reboot from console. While I can continue to run on the USB dongle, I would rather figure out if there is a setting or something to fix the network from quitting when using the internal NIC.

 

Thoughts on what could be the cause?

In each case I start seeing these lines in the syslog, when the Network quits:

### [PREVIOUS LINE REPEATED 2 TIMES] ###
May 17 02:08:31 Thor kernel: net_ratelimit: 9 callbacks suppressed
May 17 02:08:31 Thor kernel: r8169 0000:08:00.0 eth0: rtl_chipcmd_cond == 1 (loop: 100, delay: 100).
May 17 02:08:31 Thor kernel: r8169 0000:08:00.0 eth0: rtl_ephyar_cond == 1 (loop: 100, delay: 10).
### [PREVIOUS LINE REPEATED 5 TIMES] ###
May 17 02:08:31 Thor kernel: r8169 0000:08:00.0 eth0: rtl_eriar_cond == 1 (loop: 100, delay: 100).
### [PREVIOUS LINE REPEATED 2 TIMES] ###
May 17 02:10:10 Thor kernel: net_ratelimit: 9 callbacks suppressed
May 17 02:10:10 Thor kernel: r8169 0000:08:00.0 eth0: rtl_chipcmd_cond == 1 (loop: 100, delay: 100).
May 17 02:10:10 Thor kernel: r8169 0000:08:00.0 eth0: rtl_ephyar_cond == 1 (loop: 100, delay: 10).

thor-diagnostics-20230517-0652.zip

[jbat66]

This has been driving me crazy. I purchased two different TP-Link NIC's and still have the problem. I thought maybe it was the internal NIC. Looks like it is just a problem in the Unraid code. Maybe it is the fact the on board NIC and the two I purchased are running RealTech chipsets. The funny thing is that Unraid worked great for years on this server with the internal NIC. I was away for a few months, and when I got back to the server it started having the problem. So it wasn't because of an update, unless it was a plugin update that caused it. Frustrating. Every time I pull diag, I get the same lines in syslog.

 

[JorgeB]

While Realtek NICs are not recommend they work for most, don't remember seeing a similar issue, but try with an Intel NIC if posssible.

[jbat66]

14 hours ago, JorgeB said:

While Realtek NICs are not recommend they work for most, don't remember seeing a similar issue, but try with an Intel NIC if posssible.

As a matter of fact, I just installed in Intel NIC 2 min ago, and I'm waiting for reboot. We will see how long it stays up. Thanks for the info.

[jbat66]

OK, so changing to Intel NIC, did not work.

I upgraded to 6.12 RC6 - thinking the updated Kernel would help - that did not work.

I'm really out of ideas. I couldn't download a diag this time, because with 6.12 RC6, I couldn't get the console screen to come up.

 

Heck even a command to restart networking form command line would be helpful, as I could setup a cron job to restart it when it can not longer ping the gateway. I have one that will reboot the server when it can't ping the gateway but that is hard on the system, because half the time I have to rebuild a drive.

[itimpi]

3 hours ago, jbat66 said:

because half the time I have to rebuild a drive.

This sounds like an issue that needs resolving anyway.   I would never expect a reboot to mean a drive needs rebuilding if things are working as expected.