Bug#504467: Bug unreproducible

David Baron d_baron at 012.net.il
Sun Nov 9 18:50:43 UTC 2008

On Sunday 09 November 2008 17:12:48 Christian Perrier wrote:
> Quoting David Baron (d_baron at 012.net.il):
> > As I posted, the problem is NOT in the boinc-app-seti package at all.
> >
> > The start-stop-daemon for boinc (boinc-client) was producing a pid file
> > containing a different pid number than the actual boinc pid! This would
> > cause the distribution /etc/init.d/boinc_client fails to stop the boinc
> > and this causes all the installation problems for seti.
> >
> > I changed that script to simply NOT test the pid number but killall boinc
> > if any such pid file exists. Boinc-app-seti will now install.
> >
> > This but, degraded as appropriate belong to either boinc (boinc_client)
> > or start-stop-daemon.
> Well, in such case, the bug must be reassigned... probably to boinc.
> I would also downgrade it to important.
> I don't think that using killall in the init script is really
> policy-compliant. I wouldn't recommend doing this.

True. Using start-stop-daemon without the errant pid file (should) has the 
same action, however, according to the docs.

More information about the pkg-boinc-devel mailing list