Liane Praza writes:
> > A little more clarification on that: if we could make it so that no
> > package ever used a {pre,post}{install,remove} or class action script,
> > and that all of them used transient SMF service, so that package
> > installs became nothing more than cpio blasting bits, then I might get
> > more interested.
>
> That's what I'd like to see as a strategy, though I've seen no complete
> proposal yet. Lots of folks have been knocking about in support of a
> similar idea. Tony's suggested an 'upgrade' method on this alias, which
> may be one possible implementation of such an idea, though not the only one.
It's an interesting idea, but someone has to figure out what it all
means for the existing packaging ABI and for complicating features
such as diskless.
If it results in paring down the number of obscure features we
support, though, that'd actually be a good side-effect.
> (From an SMF point of view, I'm trying to make sure we're working to
> preemptively resolve things which would make that strategy difficult.)
>
> JDS's postrun is a similar idea, but gloms it all into a single service
> rather than allowing expensive or long-running actions to be parallelized.
It's a shame that the notable examples are sources of irritation.
--
James Carlson, Solaris Networking <james.d.carlson at sun.com>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
