On Fri, Nov 01, 2013 at 05:39:15PM +0000, Ian Jackson wrote:
> Steve Langasek writes ("Bug#727708: Value of reading other's position statements [was: systemd vs. whatever]"):
> > I agree. It would still require some fiddling to make 'expect stop' work
> > together with strace anyway, since upstart only cares about SIGSTOP raised
> > by upstart's child process, not by the grandchild; so if you actually need
> > upstart to know non-racily when the service is started you would need the
> > process under trace to SIGSTOP its own parent. Not elegant, but possible.
> Perhaps upstart could be made somehow to spawn strace -p at the
> appropriate moment.
> stracing daemon startup (and indeed anything else which seems to be
> malfunctioning) is a powerful tool that the competent but desperate
> sysadmin will reach for in many situations. Making it difficult is a
> distinct downside for any init replacement.
Sure; but I think the difficulty here is overstated. strace and upstart
service readiness have adverse interactions with one another, but when
you're debugging a daemon you are unlikely to be doing so under conditions
where you are simultaneously worried about upstart service readiness races.
I agree with all of the technical critiques here, I just don't see that this
relatively minor issue, which can be documented and worked around (and
ultimately, fixed upstream), is something that should be driving Debian's
choice of init system.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
slangasek@ubuntu.com vorlon@debian.org
Attachment:
signature.asc
Description: Digital signature