ubuntu-phone team mailing list archive

[Steve Langasek <steve.langasek@xxxxxxxxxxxxx>]

On Fri, Jan 10, 2014 at 11:24:52PM +0100, Michał Sawicz wrote:
> On 10.01.2014 20:04, Steve Langasek wrote:
> >I don't see how this analysis can be correct.  upstart receives no signal
> >from the kernel that the process has died until after the core handler is
> >finished, and if the unity8 process has died with a segfault there's no
> >process for upstart to kill anyway - so the kill timeout can't be related
> >here.

> >A syslog from the time of one of these failures, with 'initctl log-priority
> >info', may be enlightening.

> The thing here is that the process segfaults on exit - after it
> receives SIGTERM from upstart, which then only waits $kill_timeout
> before SIGKILLing it, at which point apport is interrupted.

Oh, ok - yes, dying with SIGSEGV after receiving SIGTERM from upstart would
certainly trigger that behavior.

Is the SIGTERM being sent because the user session is exiting?  If so,
overriding the kill timeout may not be enough, because upstart will also
hard-kill any lingering processes after a separate timeout in order to avoid
being hard-killed /itself/ by lightdm and leaving spare processes behind.

I don't think there's currently any way to configure upstart's session-end
timeout.  We should probably support configuring this via the init
commandline, so that display managers like lightdm can control it
externally.

-- 
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@xxxxxxxxxx                                     vorlon@xxxxxxxxxx

Attachment: signature.asc
Description: Digital signature