Thanks Danek for clarifying. But one last question. You state that the service runs asynchronously. But is this also the case when uninstalling? If so, could it happen that the service itself is being uninstalled before it has been run?
Joachim Danek Duvall wrote: > On Wed, Jan 14, 2009 at 04:59:54PM +0100, Jochen wrote: > > >> Do you agree that the scenario where the service is not registered is a bug? >> > > No. There's a bug open that should alleviate the problem you're having: > > 5597 service update failure during install > > This causes an install failure if the service we try to restart doesn't > exist. We should ignore that failure. Once we do, then when your service > is imported, it will start up (because you've enabled the service by > default) and do all the importing you want to do. This is precisely the > way it would work if you installed onto a non-live image, and then booted > that image. > > >> Is there a way so that the installation of a service occurs synchronously? >> > > No. > > >> Assuming there are several dictionaries, where each file triggers the >> service, would this happen synchronously? That is, only if the start >> method returned then the service is again restarted on behalve of the >> next dictionary. >> > > For any run of "pkg install" or the GUI equivalent, each unique service > FMRI is restarted exactly once, so the fact that the service isn't > restarted in a synchronous fashion shouldn't be an issue. > > Danek > _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
