[lug] Debugging a startup problem

Lee Woodworth blug-mail at duboulder.com
Wed Jul 11 12:15:27 MDT 2018


If the software hasn't changed, then it suggests a
config changed. Do you happen to a have a git repo
for /etc that would show what has changed?

Maybe ahavi, systemd and udev are competing for/or
resposnding to the kernel network device events. Do you
need avahi?

For my kernel, the first messages are from the device driver,
then udev renames the interface and the device driver reports
the name change. Looks like you have 3 things messing with the
interface: init (systemd?), udev (systemd-udev) and avahi.


On 07/11/2018 10:53 AM, Jed S. Baer wrote:
> On Wed, 11 Jul 2018 10:32:25 -0600
> Michael J. Hammel wrote:
> 
>> WAG time:
>>
>> Seems to me that the link ready messages from the kernel represent
>> state change that are not happening without being poked by higher layer
>> software.  That would suggest that the driver is not noting state
>> changes on its own or that, possible due to the if name change, the
>> events are not associated with eth3 but with eth0 at boot time so
>> network manager doesn't know to bring up that interface.
>>
>> I blame systemd.  Because I can.
> 
> I'd like to, but I think in this case, no. OTOH, did Poettering write
> network-manager too? ;-)
> 
> This is new behavior. No kernel change, no underlying software change.
> I'd try to blame the hardware, but it happens on a soft reboot as well as
> a full power off restart. Heck, even my USB ports are back now, after a
> full power cycle. I don't expect _that_ to last, but for now, I have
> nothing to look at, as far as I can tell.
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> 



More information about the LUG mailing list