Justin Moore writes:
Several people have raised this on the MythTV lists and provided
feedback to
the systemd developers. Last I heard their response was "this isn't broken
because we don't know that it's a virtual network interface and we can't wait
for ALL the network interfaces to come up because .... reasons. Closed:
Working As Intended."
The work-around was to buy a UPS for that machine because the #1 cause of
reboots was flaky power in our neighborhood. No reboots ~> fewer chances for
systemd to do the wrong thing.
Sadly I think the power company will fix the distribution infrastructure in
our neighborhood before systemd gets this working correctly.
I have no doubt that there are people who find systemd useful for one reason
or another, and they are so thrilled with joy is that they will choose to
willingly defend its merits in public.
But I can't name one off the top of my head. The systemd advocates that I've
read – in my experience they were either directly, or closely, involved
with the project.
The converse is true, for the other side. Of all the spilled ink that I
read, I can't give one name of anyone who's involved, in any way, with
anything that might be considered as competing with systemd's mind share, in
any way, shape, or matter. I see people from all walks of life, find
something or other that's fundamentally wrong with it. Either its
fundamental technical flaws, or the fact that it takes ten bleeping seconds
to figure out if one service is running, or not. The only thing they ever
have in common is that if they raise it through the official channels, the
response they usually get is not exactly professional.
For example, let's take your specific case here. I note that the follow-up
response to the article I cited made some kind of tenuous comparison with
the Linux kernel, in some form or fashion, that I found quite hard to
follow. I can't help but recall that the Linux kernel maintains blacklists
of devices, by their USB or ATAPI id, of some sort, when those specific
devices fail to implement some key aspect of the relevant hardware protocol.
The blacklisted devices are fully supported, except for some obscure
featureset for which there's a fallback alternate implementation, which is
used.
I even recall one newsworthy incident several years ago. Chipmaker FTDI
makes USB bridge chips, and a Windows driver for them. Naturally, a bunch of
cheap knockoffs of the hardware flooded the market. Since Windows already
had a driver in the base OS, it was plug and play with the clones. Then,
FTDI released a driver update that bricked the cheap clones, by
intentionally wiping the bridge chips' NVRAM, and setting their USB
manufacturer id to 0000. Anyway, some time later the Linux kernel was
patched to recognize the phony manufacturer id, and fully support it. The
incident you described reminded me of that episode, for some reason.
Anyway, back to your situation: the responsible, and user friendly, thing to
do here would also be to maintain a known blacklist of virtual network
interfaces that are not true network interfaces, and ignore them. Just like
the Linux kernel would do in a comparable situation. I'm sure that this is
eminently possible. That's what I would do. Then, the main functionality of
systemd will work as intended. This is the proper way to go, instead of the
stock "too bad, so sad, your system is going to be bleeked up because we are
smarter, and we know better" response.
I have complete faith that this is going to correct itself, over time. It
won't happen soon, but these things typically work themselves out in long
term. Currently, systemd has mindshare in most Linux distros. But that can
always change, and I think it will. And then, systemd will eventually join
XFree86 on the ash heap of history; and it will be interesting to see if
systemd will take Red Hat down with it. Red Hat has bet its farm on systemd;
and it remains to be seen how far they will ride the bomb (Google it).