Bug 838940 – sys-apps/systemd-utils-250.4 the network failed to start after reboot

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description email200202 2022-04-17 14:12:42 UTC]

After upgrading to sys-apps/systemd-utils-250.4, the network failed to start. I have "network-sandbox" enabled.

I had to mask the following packages to get the network starting again:

>virtual/udev-217-r3
>virtual/libudev-232-r5
>sys-fs/udev-249.9
>sys-apps/systemd-tmpfiles-249.9
>virtual/tmpfiles-0-r1

Reproducible: Always

Steps to Reproduce:
1.emerge --sync
2.emerge -auvDN world
3.reboot
Actual Results:  
Apr 17 17:28:11 [/etc/init.d/net.eth0] ERROR: interface eth0 does not exist
Apr 17 17:28:11 [/etc/init.d/net.eth0] Ensure that you have loaded the correct kernel module for your hardware
Apr 17 17:28:11 [/etc/init.d/net.eth0] ERROR: net.eth0 failed to start


Expected Results:  
The kernel modules are loaded and the network starts

[Comment 1 Sam James 2022-04-17 14:17:27 UTC]

Can you include emerge --info? Also, what were you using before? eudev or sys-fs/udev?

[Comment 2 email200202 2022-04-17 14:20:17 UTC]

Created attachment 771110 [details]
# emerge --info '=sys-apps/systemd-utils-250.4::gentoo'

[Comment 3 Ionen Wolkens 2022-04-17 14:50:07 UTC]

Unsure if it's related to what's happening here, but may want to try -r1 (>=systemd-utils-250.4-r1) when you get it in emerge --sync

https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=c2318993fac0cdc54dddc3892816dc7992bb0902

[Comment 4 Mike Gilbert 2022-04-17 15:04:56 UTC]

udev probably renamed your eth0 interface to something else.

What method are you using to disable network interface renaming?

[Comment 5 Benjamin Réveillé 2022-04-17 15:10:52 UTC]

I think this is a kernel module autoload issue in udev-250/systemd-utils-250.4

I say this because on my side amdgpu was not autoloading...

Got it working again by downgrading udev/systemd-utils which was possible by by masking the following packages in /etc/portage/packages.mask
=sys-fs/udev-250
=virtual/udev-217-r4
=virtual/libudev-232-r6
=virtual/tmpfiles-0-r2
=sys-apps/systemd-tmpfiles-250
=sys-apps/systemd-utils-250.4

[Comment 6 Sam James 2022-04-17 15:23:36 UTC]

(In reply to Benjamin Réveillé from comment #5)
> I think this is a kernel module autoload issue in
> udev-250/systemd-utils-250.4
> 

What about >= -r1? See https://bugs.gentoo.org/838940#c3.

[Comment 7 Benjamin Réveillé 2022-04-17 16:10:38 UTC]

(In reply to Sam James from comment #6)
> (In reply to Benjamin Réveillé from comment #5)
> > I think this is a kernel module autoload issue in
> > udev-250/systemd-utils-250.4
> > 
> 
> What about >= -r1? See https://bugs.gentoo.org/838940#c3.

systemd-utils-250.4-r1 fixed it for me and kernel module autoloading (for amdgpu).

If I recall correctly, systemd-utils-250.4 and udev-250 where both installed previously... whereas now only systemd-utils-250.4-r1 is installed

[Comment 8 Mike Gilbert 2022-04-17 23:19:49 UTC]

Please re-open the bug if you still experience the problem with >=sys-apps/systemd-utils-250.4-r1.