Re: Bug#727708: tech-ctte: Decide which init system to default to in Debian.
Brian May <brian@microcomaustralia.com.au> writes:
> On 29 October 2013 12:21, Russ Allbery <rra@debian.org> wrote:
>> In other words, I don't think it would make any sense at all to
>> standardize on upstart or systemd and then ask people to continue to
>> write init scripts in the long run (transition issues aside). Getting
>> rid of init scripts is not the whole point, but it's a huge part of it.
> My understanding is that init scripts will still be required for FreeBSD
> and The Hurd.
I would not assume that. At least, I personally don't think that
switching to upstart or systemd as a default but requiring that everyone
provide both files for that system and init scripts for Hurd and kFreeBSD
to be a good outcome, since I don't think that will be something at which
Debian will be successful.
There are various other options, including not changing away from sysvinit
or someone porting the necessary support to Hurd and kFreeBSD. Or, of
course, dropping Hurd and kFreeBSD, although I'm sure that no one wants
that outcome.
> Which is unfortunate. I often encounter init scripts that have race
> conditions or are otherwise broken. Restarting a simple daemon becomes a
> challenge to debug the supposedly good init script first.
> If the majority switched to systemd, then it would be up to the minority
> to maintain the initd scripts. I don't see this working very well.
Exactly.
Many of the existing init scripts are of poor quality in edge case
behavior precisely because writing a robust init script is exceedingly
difficult. I think we have to ask whether this is something that we
should expect Debian packagers to continue to spend their time on.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: