3 ms·
No. I think the linked article is pretty clear where the problems are. It's basically assumptions made by systemd that are not always true and from the looks of
by nisa 12y ago
No. I think the linked article is pretty clear where the problems are. It's basically assumptions made by systemd that are not always true and from the looks of it a possible bug regarding the ordering of service files and reaction to events. Besides that broken or unusable console output without plymouth (I can't really tell from the article if systemd is responsible for that).
That's all fine and possible expected.. my only concern is just that this appears to be very difficult to debug. The author is as knowledgeable as you can hope for in a Linux user and is unable to get meaningful debug output.
All of these things are IMHO legitimate concerns that should be addressed by the systemd community or the Debian systemd packagers.
- digi_owl 12y ago>It's basically assumptions made by systemd that are not always true Surprise surprise. Assume nothing, test everything is a old mantra in various engineering circles iirc.