Stephen Hahn wrote: >* Casper.Dik at Sun.COM <Casper.Dik at Sun.COM> [2007-10-29 20:38]: > > >>>(I'd like to completely eliminate the possibility of setting the two >>>differently, but I suspect that some scenarios require it.) >>> >>> >>It's funny that the one thing I really would like to have: enable at >>next boot, is currently not possible and not possible in the proposal >>on the table either. Or is that such an unnatural thing to want. >> >> > > It's not unnatural--although it's inappropriate for most services, and > generally discouraged--but there is a way to do it. >
Why is it inappropriate for most services? And what is "most services" in this context? >(Some of the > postinstall scripts know how; it's accomplished by manipulating > general/enabled and general_ovr/enabled in a coordinated way.) So, I > think the question is "when can someone do an enable-for-next-boot > RFE for svcadm(1M)?" > > Whenever I think about this for more than a few minutes, it seems that > I would rather have the few services that really need this return an > appropriate exit status if their start method is run before their > boot-related state is readied. But that might actually be > unnatural... > > I think I understand what you're saying here, but I'm not sure... Would you mind expanding on this a little more? Could you give an example of the problem you're thinking of and how the special exit status helps? Thanks, Darren
