I just attempted to upgrade to 6.2 (finally noticed it's stable), but I ran into a problem similar to that in this thread:

https://lime-technology.com/forum/index.php?topic=50215.msg481963#msg481963

 

After reboot into 6.2, I was unable to access my server.  Upon logging in locally I was greeted with the message "Device "eth0" does not exist."

 

It appears from the logs that eth0 is getting renamed, eventually to eth120.

 

The difference from the other thread being, my motherboard (Asus P8Q77-M/CSM) has only one physical NIC, I can't disable an "extra" NIC.

 

I've since rolled back to my prior install and all is well again, but I'd like to upgrade to 6.2.  I captured the diagnostics and attached the syslog.  If there's anything I can do on my end to debug or fix this it would be great.

 

-edit

 

Well, I do see two NICs in the logs, e1000e and e1000 with different drivers.  I suspect that's related to AMT, but there's only one RJ45 port on the board.

syslog.txt

Since no one else has come forward with any help, let me make a suggestion.  Install a virgin copy of unRAID on a new drive and see if that boots properly.  If it does, it says that there is some incompatibility between ver 6.2 and some plugin, Docker or VM from your old setup.

 

Second item, you did not tell us much about the 'old' setup.  What version of unRAID were you running?  What plugins, Dockers , and VM's were you using?  Did you read and follow the upgrade instructions in the release threads for all of the versions that you might have skip over? 

 

Third, if you can reinstall version 6.2, use a monitor and keyboard, login and type  diagnostics  at the command line.  That will write the diagnostics file to the Flash Drive.  Attach that file to a new post.

You have 2 NICs, one PCIe and another PCI.

 

Remove one of them or use the new network manager to choose which one is eth0.

 

Edit: One of them is the onboard one, probably the PCIe

Well, I'm both quite embarrassed, and somewhat "vindicated". 

 

I don't know what I was looking at last night, I looked for ethernet ports and my brain only processed one.  But yes, I did have  PCI card installed (embarrassed).

 

So I removed that and tried the update again.  Unfortunately the same problem returned, eth0 is still getting renamed (I feel a bit vindicated that it wasn't entirely me being blind and not seeing/removing the second NIC last night).  I had actually captured the whole diagnostics package last time, I just didn't attach it.  I've attached it below.

 

As far as plugins etc, I don't have a lot, I have one Windows 7 VM that I launch on startup, as well as Crashplan and LMS, plus sagetv (but that's not currently starting/enabled).

 

I'm kind of out of debug time for the moment so I'll have to save the fresh 6.2 install for another day.  Also I'm upgrading from 6.1.7, so as best I could tell there weren't any special upgrade instructions steps to take, other than using the Upgrade option from the plugins page and rebooting (which I did).

unraid-diagnostics-20160928-1533.zip

So I got a chance to try a fresh install, and it appeared to work OK, I managed to get an IP address and eth0 was there as expected.  I guess that means it's something with my config.

 

Is my best bet to try disabling docker/kvm before upgrading and then re-enabling them after?  Or is there some other, better option?

So I got a chance to try a fresh install, and it appeared to work OK, I managed to get an IP address and eth0 was there as expected.  I guess that means it's something with my config.

 

Is my best bet to try disabling docker/kvm before upgrading and then re-enabling them after?  Or is there some other, better option?

Yes I think that would work.  Also look in the XML for your VM to see if you are passing through a nic.  If you are then that is what is probably causing your problems as that nic was now being used as your unRAID nic after the upgrade.

 

 

Specifically I saw the following:

e1000e 0000:00:19.0 eth0: removed PHC

which if I remember correctly I always saw when I passed through a nic that wasn't stubbed in syslinux.cfg.

 

 

In any case I do see in the log where unRAID changed the name to eth1:

e1000e 0000:00:19.0 eth1: renamed from eth0

Also I'm upgrading from 6.1.7, so as best I could tell there weren't any special upgrade instructions steps to take, other than using the Upgrade option from the plugins page and rebooting (which I did).

 

You might want to review the first two posts in the release thread as well:

  https://lime-technology.com/forum/index.php?topic=51874.0

There is a section on troubleshooting networking

Thanks, that might be helpful, when I get some debugging time again, I think I'll have to try deleting the network.conf.

 

Yes I think that would work.  Also look in the XML for your VM to see if you are passing through a nic.  If you are then that is what is probably causing your problems as that nic was now being used as your unRAID nic after the upgrade.

 

 

Specifically I saw the following:

e1000e 0000:00:19.0 eth0: removed PHC

which if I remember correctly I always saw when I passed through a nic that wasn't stubbed in syslinux.cfg.

 

 

In any case I do see in the log where unRAID changed the name to eth1:

e1000e 0000:00:19.0 eth1: renamed from eth0

 

Yeah, I saw that, but no I don't have a nic passed through:

<interface type='bridge'>
      <mac address='##:##:##:##:##:##'/>
      <source bridge='br0'/>
      <target dev='vnet0'/>
      <model type='virtio'/>
      <alias name='net0'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/>
    </interface>

FYI, I did get 6.2 to update and load my network correctly by deleting (renaming) my network.cfg file