View previous topic :: View next topic |
Author |
Message |
Spear n00b
Joined: 08 Mar 2004 Posts: 65
|
Posted: Mon Oct 13, 2014 4:43 pm Post subject: <Solved> Netmount depends upon an interface that's gon |
|
|
So on a fresh install, I've reached a point where netmount tells it can't start because net.eth0 can't be started.
Which is odd since this box never had a net.eth0 to begin with.
The two adapters start as eth0 and eth1 and are renamed to enp4s0 and enp7s0. These two interfaces work just fine, they use the sky2 driver and if I manually start their init scripts they both come up and function just fine.
But netmount chokes complaining about an interface that this machine never had configured. Poking around should that /etc/conf.d/netmount only has rc_need set for the net config file and that works fine.
I can't find any other reference to an interface of net.eth0 in configuration, just references to it as an example in comments. There's no script or symlink for it, there's no reference to it in the runlevels.
So where is netmount getting the idea that there should be one?
Last edited by Spear on Tue Oct 14, 2014 5:28 pm; edited 1 time in total |
|
Back to top |
|
|
Roman_Gruber Advocate
Joined: 03 Oct 2006 Posts: 3846 Location: Austro Bavaria
|
Posted: Mon Oct 13, 2014 6:13 pm Post subject: |
|
|
there is a kernel flag, which you can add to the knerel invocation line which gives you the old behaviour
something like net-ifnames ... check your search engine of choice for it.
the new naming scheme will break things and thats why i use this kernel parameter, no hassle. |
|
Back to top |
|
|
Spear n00b
Joined: 08 Mar 2004 Posts: 65
|
Posted: Tue Oct 14, 2014 5:31 pm Post subject: |
|
|
I've figured it out based on clues from this thread:
https://forums.gentoo.org/viewtopic-t-997600-view-next.html?sid=a2401cfe3750c0e7efaa971adc1ce036
When I was assembling the machine I set the date wrong in the BIOS, i.e. normal style when the BIOS expected the American style (mm/dd).
This was corrected later after the install was partially done. This lead to clock skew on the syslog-ng entry meaning the rc-update commands weren't taking effect but weren't making me aware at the time. It was the rc-update -u command that would reveal the clock skew complaint.
A simple touch on /etc/conf.d/syslog-ng fixed it. |
|
Back to top |
|
|
|