Bug#781886: qcontrol failure to start on boot sometimes (jessie, systemd?)

[Ian Campbell]

On Thu, 2015-04-16 at 10:28 +0200, reportbug wrote:
> On Wed, Apr 15, 2015 at 19:19:24 +0100, Ian Campbell wrote:
> > @reportbug (sorry I don't seem to know your real name), could you try
> > adding a call to "udev settle" to the qcontrold initscript, in the start
> > case right before the daemon is launched (I can be more specific if you
> > need).
> 
> Hi. It pains to say that I'm currently remote from my qnap machines, and
> also, because of a lack of a (serial) console I am extremely wary of
> fiddling with boot scripts on these machines, since they are doing useful
> work and holding useful data. I wish I could be of more help, but sadly I
> don't have a spare machine for debugging work.

No problem, I've not been able to reproduce myself but perhaps I just
need to keep rebooting some more.
> 
> (And sorry for just using a one-off email address, but posts to the 
> Debian bugtracker seems to be an incredibly strong magnet for spammers 
> and email harvesters)

Understood all to well :-/

> Thanks again for your time and your work. :)

My pleasure.

> For anyone testing, I can say that the problem happens only occasionally. I
> think once out of the ~4 reboots I have performed so far.

Thanks, that's a useful datapoint, I've rebooted ~4 times too without
seeing it, but I'll try a couple of orders of magnitude more attempts
before I try some more drastic hacks to reproduce.

BTW, when this fails the system still booted and you could login and fix
it, i.e. it didn't stall the boot or anything, is that right? If that is
the case a release note might be the most plausible action at this
point.

Ian.