Bug 1142800

Summary: systemd-networkd with kernel 5.2 don't up network interfaces
Product: [openSUSE] openSUSE Tumbleweed Reporter: Илья Индиго <ilya>
Component: NetworkAssignee: Franck Bui <fbui>
Status: RESOLVED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: dheidler, thomas.blume, tux93
Version: CurrentFlags: fbui: needinfo? (thomas.blume)
Target Milestone: ---   
Hardware: All   
OS: openSUSE Factory   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Илья Индиго 2019-07-25 09:51:56 UTC 
The network is configured via systemd-networkd.
After update Tumbleweed to Kernel 5.2.1, I got a dead network.
At the same time, network and systemd-networkd are running and restarting them does not give anything.
The network is activated on only manually.
sudo ip link set enp3s0 up

If I updated remotely and restarted the computer, I would get a dead server.

Maybe openQA should add tests for systemd-netword?
Comment 1 Илья Индиго 2019-07-25 09:53:40 UTC 
https://github.com/systemd/systemd/issues/12784
Comment 2 Илья Индиго 2019-07-25 12:17:09 UTC 
SR https://build.opensuse.org/request/show/718032 fix network activation on startup.
Comment 3 Franck Bui 2019-07-25 12:51:17 UTC 
Thanks for the report. As you already spotted, the fix has been submitted and should be part of the next update.

Hence closing.
Comment 4 Илья Индиго 2019-07-25 13:44:48 UTC 
What about to prevent this in the future, and do QA tests for this?
Or does openQA test only what is by default, and those who use the systemd-netword do it at their own peril and risk?
Comment 5 Franck Bui 2019-07-25 13:58:23 UTC 
Can you ask to upstream instead ?
Comment 6 Илья Индиго 2019-07-25 15:15:40 UTC 
It’s useless to ask Lenard "Notabug" Pottering upstream, there are at least a dozen of such bugs.
As it seems to me, the distribution kit in its QA should be engaged in checking the basic operability of the distribution kit.
I do not know how openQA is organized here, and how and who to offer?
Not all in a row to ask dimstar by email.
Comment 7 Илья Индиго 2019-07-25 15:23:21 UTC 
Moreover, I think that they even tested, but not with kernel 5.2, since they simply did not have it and could not do it.
Comment 8 Franck Bui 2019-07-25 15:27:57 UTC 
Thomas, do you think this should worth to be check in openqa ?

IMO this should be checked by upstream so such issues are caught and fixed by them directly.
Comment 9 Franck Bui 2019-07-25 15:30:17 UTC 
That said maybe it's already tested, it's just that kernel 5.2 has not yet been submitted and tested...

One more reason for upstream to test it I would say.
Comment 10 Franck Bui 2019-07-25 15:33:00 UTC 
It looks like I missed your comment #7.
Comment 11 Dominik Heidler 2019-07-26 13:30:49 UTC 
(In reply to Franck Bui from comment #8)
> Thomas, do you think this should worth to be check in openqa ?
> 
> IMO this should be checked by upstream so such issues are caught and fixed
> by them directly.

This actually made openQA tests fail: https://bugzilla.opensuse.org/show_bug.cgi?id=1141597
https://openqa.opensuse.org/tests/990747
Comment 12 Franck Bui 2019-07-26 15:08:21 UTC 
(In reply to Dominik Heidler from comment #11)
> This actually made openQA tests fail:
> https://bugzilla.opensuse.org/show_bug.cgi?id=1141597
> https://openqa.opensuse.org/tests/990747

Nope, again this one is a different issue.
Comment 13 Franck Bui 2019-07-29 07:22:27 UTC 
*** Bug 1142901 has been marked as a duplicate of this bug. ***