I am having an issue with eth0 on my unraid server. This setup has been running flawlessly for a few weeks, brand new components, all of a sudden the server was unresponsive and after reboot the nic will almost immediately fail again. I have attached the diagnostics file for assistance. I just need to know if i need to RMA the MB or if this is a software issue. 

 

Edit: I have figured out what causes it, i have no idea why. See post below. TLDR: Hassio VM running motioneye inside of docker causes the conflict. 

 

MB: Asrock z390m pro 4

CPU: G5420 

onboard nic: I219-v

 

 

May 30 00:39:20 Tower kernel: e1000e 0000:00:1f.6 eth0: Detected Hardware Unit Hang:
May 30 00:39:20 Tower kernel:  TDH                  <15>
May 30 00:39:20 Tower kernel:  TDT                  <59>
May 30 00:39:20 Tower kernel:  next_to_use          <59>
May 30 00:39:20 Tower kernel:  next_to_clean        <14>
May 30 00:39:20 Tower kernel: buffer_info[next_to_clean]:
May 30 00:39:20 Tower kernel:  time_stamp           <ffff348b>
May 30 00:39:20 Tower kernel:  next_to_watch        <15>
May 30 00:39:20 Tower kernel:  jiffies              <ffff4b40>
May 30 00:39:20 Tower kernel:  next_to_watch.status <0>
May 30 00:39:20 Tower kernel: MAC Status             <40080083>
May 30 00:39:20 Tower kernel: PHY Status             <796d>
May 30 00:39:20 Tower kernel: PHY 1000BASE-T Status  <3800>
May 30 00:39:20 Tower kernel: PHY Extended Status    <3000>
May 30 00:39:20 Tower kernel: PCI Status             <10>
May 30 00:39:21 Tower kernel: e1000e 0000:00:1f.6 eth0: Reset adapter unexpectedly
May 30 00:39:21 Tower kernel: bond0: link status definitely down for interface eth0, disabling it
May 30 00:39:21 Tower kernel: device eth0 left promiscuous mode
May 30 00:39:21 Tower kernel: bond0: now running without any active interface!
May 30 00:39:21 Tower kernel: br0: port 1(bond0) entered disabled state
May 30 00:39:25 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
May 30 00:39:25 Tower kernel: bond0: link status definitely up for interface eth0, 1000 Mbps full duplex
May 30 00:39:25 Tower kernel: bond0: making interface eth0 the new active one
May 30 00:39:25 Tower kernel: device eth0 entered promiscuous mode
May 30 00:39:25 Tower kernel: bond0: first active interface up!
May 30 00:39:25 Tower kernel: br0: port 1(bond0) entered blocking state
May 30 00:39:25 Tower kernel: br0: port 1(bond0) entered forwarding state

tower-diagnostics-20200530-0404.zip

Edited May 31, 20206 yr by shanetron

I have since added a usb RTL8152 based nic. This has stabilized the system, however it appears to only use the usb interface, and will not use the on board nic at all now unless i unplug the USB interface, the bonding settings are set for backup.  

 

I re-flashed the usb disk and recreated the array to see if a new install would solve the issue, and it came back almost immediately. Sounds to me like a hardware issue since it worked perfectly for almost 3 weeks. Is there an issue with the Intel I219-v that i am unaware of?

Edited May 30, 20206 yr by shanetron

I have gotten close to figuring it out, but i do not understand why. 

 

Through all sorts of process of elimination, I have discovered that "motion eye" running in docker, inside of a VM of hassio(home assistant), when it retransmits video, that is what causes the ethernet adapter to become unstable. Originally i thought it would be a port conflict of some kind since the retransmit is on 8081, i tried 8082 and 8083, neither stopped the issue. 

 

Does anyone have any idea what could be causing this conflict? I am assuming is must be an issue with the vir ethernet adapter that is created for the vm.