[arch-general] testing/systemd 191-1 failed to boot

Kevin Chadwick ma1l1ists at yahoo.co.uk
Mon Sep 24 05:55:19 EDT 2012

> ... but i think we all agree openssl is worth keeping around, yes?
> such exceedingly high standards of "quality" and "stability" are
> simply unrealistic.

That all depends on your time to implement new functionality

> systemd works pretty awesome for everything i've thrown at it, and
> minus a few kinks here and there, flawless as well.  of course, YMMV,
> but i'd surmise 98% of those having difficulties are simply making
> some small error ... systemd, IME, pretty much does what you'd expect.

As 90% of systemd (especially the useful parts) is already catered for
by other software such as all of the so called and incorrectly
documented security benefits, this time to functionality just causes
problems unnecessarily. Systemd has merits but as it is fundamentally
flawed without a good enough reason because of such a large pid1 and
dependence, there is no point in polishing it or me ever spending any
time using it.

> btw, per Allan's link, the guy who made himself more useful than any
> of us -- ie, by reporting (gasp!) the bug and including a patch --
> concluded with this little gem:
> "[...]
> PS: I've been using systemd since v188 (w/o sysvinit) for
> desktop/laptops (and even in ramroots) and love it. The unit file
> syntax is nice and intuitive and things mostly work great
> out-of-the-box. This, is after several years of using Upstart, which
> worked adequately, but still was not as flexible as it should've been,
> and a was bit fragile.."
> ... just some food-for-thought is all.

We will ignore the noob and migrater barriers as this seems to apply to
a lot of primarily linux software with more than a single binary (google
required). Consider mounting a logical volume without prior reference
by looking at the lvm page. The link between lvm and lvchange to enable
mounting simply isn't there. I guess I'm used to the practicality of
the excellent OpenBSD documentation, which is one of their goals


'Write programs that do one thing and do it well. Write programs to work
together. Write programs to handle text streams, because that is a
universal interface'

(Doug McIlroy)

More information about the arch-general mailing list