http://defect.opensolaris.org/bz/show_bug.cgi?id=10805



--- Comment #5 from Renee Danson Sommerfeld <renee.danson at sun.com> 
2009-08-20 17:26:40 UTC ---
(In reply to comment #4)
> That seems reasonable, but the problem is the order in which we carry out
> operations such as commit and destroy is:
> 
> 1) update the configuration in netcfgd
> 2) update the state in nwamd
> 
> This order is needed for commit-type events since nwamd re-reads the
> configuration for most operations (e.g. commit triggering a refresh). If we
> accept that this ordering is necessary, we are left in the awkward situation
> that the object is effectively already gone by the time we run the stop 
> script,
> so there's nothing to put into maintenance.
> 
> It may make sense to reverse this order for destroy events, but I'm worried
> that'll make life quite complicated. I'd also be worried that it would be
> possible to make undestroyable objects if the script or service didn't exist.

Ah, I spaced on the "destroy" part.  If the object's been destroyed, it's hard
to have state associated with it, huh!  I guess in that case the best we can do
is a log message (maybe higher than debug, so it shows up in /var/adm/messages
for the default syslog settings) noting that the stop method failed.  Sorry for
the confusion...

-- 
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.

Reply via email to