http://defect.opensolaris.org/bz/show_bug.cgi?id=14263
amaguire <alan.maguire at sun.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |alan.maguire at sun.com
--- Comment #5 from amaguire <alan.maguire at sun.com> 2010-02-03 11:54:02 UTC
---
So if I understand the problem - we want to trigger conditon checking when ENM
and location states change, but in order to do effective condition checking we
need to ensure that the location has been applied. Because we can get flurries
of refreshes, we can't easily place an upper bound on the time this will all
take. The key conditions we need to worry about here are advertised-domain and
system-domain since these are actually determined on the basis of location
application (at least sort of in the former case). So I might apply location X
with system domain a, then location Y with system-domain b, and the test might
pick up a when it expected b. Is that right?
If that's the problem I'd say what we need to do is simply trigger the
condition checks each time. If a test relies on the conditions being valid
within a certain time interval, maybe we can figure out a way of making that
test more robust (assuming nwamd does do the right thing in the end. I may have
misunderstood).
One thing I can't remember though is if there's a good reason to have location
activation conditions support system-domain at all - if not, allowing
system-domain as a condition seems circular to me since the location is the
entity that applies the system domain in the first place (at least when it's
manually specified). We already forbid locations and enm's having
self-referential activation conditions where these pertain to that
location/enm's state, this seems a similar case to me.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
