Networking in 6.8 Sorta All Messed up.

[wesman]

I am not sure it's a 6.8 issue, but I did upgrade today, and now the network is all over the place. Or maybe (and highly likely its user error)

My Motherboard has 3 Network ports, (1) 10 gbe and (2) 1 gbe.. The 10 gbe is the only thing plugged in (see image below).

However, after the upgrade, not etho is now bonded with all three, when I set the no bonding, network goes down and I have to manually connect to the server with a keyboard, mouse and monitor.  For some reason eth0 is not what it should be and I cant change it. 

I tried deleting the network cfg files in /boot/config/network*

 

but it comes back up the same way. I assume that e1000e is emulated?? since it has the same MAC address? 

 

 

 

The only way i found the fix it is, to remove the e1000e device from the network-rules.cfg  but if I make any changes, it goes right back to they way it was

 

[JorgeB]

12 hours ago, wesman said:

I assume that e1000e is emulated?

e1000e is the driver used for the other onboard NIC, both 1Gbit NIC use different drivers.

 

Can't you just make the 10GbE NIC (atlantic) eth0 and then reboot? Also make igb eth3 or eth2, it should stick.

[wesman]

At the bottom there, where I can reassign eth0.. any changes I try to make gives the error miss matched.. because there e1000e and the Atlantic both haver the same Mac address. Which is why I thought it was emulated.

[JorgeB]

I didn't noticed the duplicated MAC address, delete /boot/network.cfg, /boot/network-rules.cfg, reboot and grab and post diags after booting is complete.

[bonienl]

There is something seriously wrong with the MAC addresses of your interfaces. The OUI is set as ASrock, and can't be the same for different vendors (Intel and Atlantic). More over duplicate MAC addresses are not allowed.

 

What motherboard are you using?

Does it have BIOS settings for the ethernet settings?

Edited January 2, 2020 by bonienl

[wesman]

7 hours ago, johnnie.black said:

I didn't noticed the duplicated MAC address, delete /boot/network.cfg, /boot/network-rules.cfg, reboot and grab and post diags after booting is complete.

I did that, mentioned above, but after the reboot, I get the same thing.

The only way around is to use NANO modify the Network-Rule.cfg by removing the e1000e rule, and then it works. BUT if I make any changes after that, it goes back to this crazy state.

[wesman]

3 hours ago, bonienl said:

There is something seriously wrong with the MAC addresses of your interfaces. The OUI is set as ASrock, and can't be the same for different vendors (Intel and Atlantic). More over duplicate MAC addresses are not allowed.

 

What motherboard are you using?

Does it have BIOS settings for the ethernet settings?

The Motherboad is the ASROCK Z390 Taichi Ultimate. It has the latest BIOS 4.20.

I don’t understand what you mean by “BIOS settings for the Ethernet” 

[JorgeB]

You should post the diagnostics like asked, there might me something visible there.

[bonienl]

It seems to be a problem with this motherboard.

 

See http://forum.asrock.com/forum_posts.asp?TID=10788&title=z390-taichi-ultimate-nics-same-mac-address

 

[wesman]

14 minutes ago, johnnie.black said:

You should post the diagnostics like asked, there might me something visible there.

Attached

 

valyria-diagnostics-20200102-0719.zip

[wesman]

13 minutes ago, bonienl said:

It seems to be a problem wih this motherboard.

 

See http://forum.asrock.com/forum_posts.asp?TID=10788&title=z390-taichi-ultimate-nics-same-mac-address

 

Well, that is an interesting read... I’ll have to try that when I get home and see if it works. 

[JorgeB]

Like bonienl mentioned both onboard NICs have the same MAC:

 

Dec 31 13:38:06 Valyria kernel: igb 0000:08:00.0: eth1: (PCIe:2.5Gb/s:Width x1) 70:85:c2:c8:57:58
...
Dec 31 13:38:06 Valyria kernel: e1000e 0000:00:1f.6 eth2: (PCI Express:2.5GT/s:Width x1) 70:85:c2:c8:57:48

 

So it's a borad problem, see if the linked discussion solves the issue.

[gmihovics]

I am actually having the same issue with removing eth1 from my bond. If I remove eth1, which is not configured or connected, I lose access to the WebGUI as well. The only way to fix it is to delete the network.cfg and reboot from the terminal. After the reboot the default network settings take over again and eth1 is back in the bond and the WebGUI is working again.

 

I have a supermicro X8SIL and have verified that in /proc/net/bonding/bond0 i have two separate devices each with a different mac address.

 

Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)

Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eth0
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:25:90:60:08:b0
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 1
Permanent HW addr: 00:25:90:60:08:b1
Slave queue ID: 0

 

Edited January 11, 2020 by gmihovics

[odd]

I am also having the same dupe MAC address on all interfaces. I have a clean install of 6.8.3 on a SuperMicro X11SBA-F and I was trying to configure the bond to the switch and ran into this issue. This issue is unique to Unraid as this issue does not pop up using a different OS on the same hardware so I'm hesitant to flash the firmware of the NIC's as shown in the linked thread in this post.