On Tue, 27 Mar 2007, David Finberg wrote:
I'm answering myself since I want to piece together a few responses in
case someone else is looking for answers.
> say I have some service, and I do
> prop = scf_simple_prop_get(NULL, NULL, "config", "testval");
> if (prop) {
> if ((val = scf_simple_prop_next_boolean(prop)) != NULL)
>
> inside my service's executable. What happens if svccfg is used to change
> config/testval, the change is not committed via a refresh, but my service
> dies and restarts.
>
> What I expected was that the old, i.e. the running (svcprop -s running -p
> config/testval) state would be returned. But instead, the current state of
> the repository is returned. Is this the right behavior? And is there some
> simple explanation of why I want that in this case, since I guess my
> intuition is way off here.
I guess my intuition was better than I thought, since this behavior is
incorrect. See http://bugs.opensolaris.org/view_bug.do?bug_id=6539436
There might be a use case for returning the repository state instead of
the running snapshot, but I can't think of a good one.
This does mean that code in Solaris that uses these interfaces is probably
doing the wrong thing, like bind, rmvolmgr, etc. I see a few places where
code is checking general/enabled, I think that works since general/enabled
is a bit odd. But as a service getting my configuration properties I
would think I would essentially always want the running snapshot.
It might be nice to place this into the FAQ as well, in case developers
are tempted to use these interfaces.
-- Dave
