Karl Rossing writes:
> The cups problem should be reproducible. I just need to gather up more
> information before I submit a bug.
>
> I was seeing the svc:/application/cups/scheduler:default restart every 3
> seconds.
> [ Jun 18 13:05:27 Stopping because process dumped core. ]
> [ Jun 18 13:05:27 Executing stop method (:kill). ]
> [ Jun 18 13:05:27 Executing start method ("/usr/sbin/cupsd"). ]
> [ Jun 18 13:05:28 Method "start" exited with status 0. ]
> [ Jun 18 13:05:30 Stopping because process dumped core. ]
>
> Turns out that it was not cupsd but hpijs core'ing while trying to process
> what I think is an adobe acrobat 9.1.2(x64) job on an HP deskjet 5250. I'm
> not sure why cupds would restart if a sub process would core.
>
> So the restart of the cupsd service would cause all queued job to restart on
> the 1st page and the cycle would continue so each print job would continue to
> print out on the same 1st page.
>
> We probably printed 3000 pages today like this.
> --
> This message posted from opensolaris.org
> _______________________________________________
> smf-discuss mailing list
> smf-discuss at opensolaris.org
Karl,
By design SMF restarts services when one of its processes dumps core. The
thinking is that this is not normal behavior for a service, so perhaps the
service is no longer running properly.
As Antonello mentioned, you can change this behavior by using svccfg to set
the startd/ignore_error property of the service to "core". See the man
pages for svc.startd(1M) where the property is discussed and svccfg(1M) for
instructions on setting properties.
tom
