A space to ponder, discuss, speculate, disagree, and learn.

Expected topics include:

Technology
Politics
Defense
Philosophy
Humanism and Transhumanism
The Future

"The Tragedy of systemd" -- a rebuttal

Posted 11-17-2018 at 05:10 PM by ttk

Benno Rice gave a talk at the last BSDCAN about systemd from a BSD perspective, "The Tragedy of systemd", and I wrote a rebuttal on my blargh:

http://www.ciar.org/ttk/blog/index.c...uttal_D90F74AA

Posted in Uncategorized

Views 2969 Comments 4

« Prev Main Next »

Total Comments 4

Comments

	Bah .. 676 views here on LQ, but only 61 of you clicked through to the actual article. Next time I'll paste a copy of the entire article into LQ, not just the link.
	Posted 11-26-2018 at 01:07 PM by ttk

"If Nagios hits up a service immediately or soon after system start and starts the service, then why did we bother deferring the service start in the first place?"

This is a classic example of dis-jointed design (which is prevalent in Linux, but not so much in the BSD's). We have some feature that solves some problem, yet, the problem doesn't really exist on a higher level of the architecture. Ironically, one of the goals of systemd was to reduce the disjointed manner in which Linux distros operate.

Personally, I would love to see systemd and the logic underlying it staying as far away from the BSD's as possible, but that's just me =)

Posted 11-26-2018 at 02:54 PM by rocket357 rocket357 is offline

Quote:

Personally, I would love to see systemd and the logic underlying it staying as far away from the BSD's as possible, but that's just me =)

Not just you, sir! I for one am glad that the BSDs remain a final fallback position, should the remaining Linux distributions fall to the enemy.

For the moment, however, the situation seems fairly static. Roughly 70 Linux distributions (about 25%) do not use systemd and show no signs of starting. http://without-systemd.org/wiki/inde...ithout_systemd

The systemd team tried to make more advances by weaponizing some software dependencies. First they tried deprecating the netlink device interface for a systemd-specific interface, but device driver developers are an unruly crowd of cats and shunned the new interface. Then they tried to introduce a systemd dependency to the kernel itself with kdbus/BUS1, but the Linux devs repeatedly rejected these patches.

Either of these moves, if successful, would have swept most of the resistance into the systemd fold. But they failed, so they declared victory and pretend it's fait accompli.

So the lines of battle have remained static for a while, but who knows when they might start up again on a new front.

If the worst happens, many of us will be glad to find final refuge with the BSDs.

Posted 11-27-2018 at 01:56 PM by ttk ttk is offline

Updated 11-27-2018 at 02:05 PM by ttk

Quote:

Originally Posted by ttk

Not just you, sir!

Yeah, I did pick up on that sentiment in your blog post =)

As a long-time OpenBSD convert, I'm not really feeling any threat from systemd, but to see a FreeBSD dev discuss it as if it were the best thing since sliced bread is a bit...concerning. Not that I think Theo would ever willingly accept systemd madness into OpenBSD, though.

If ever systemd does happen to overwhelm the remaining open source operating systems out there, I'll probably retire to a remote mountain top to raise chickens and goats and never look back lol.

Posted 11-27-2018 at 02:46 PM by rocket357 rocket357 is offline