http://defect.opensolaris.org/bz/show_bug.cgi?id=12883
--- Comment #2 from James Carlson <carlsonj at workingcode.com> 2009-11-26
23:52:13 UTC ---
(In reply to comment #0)
> To solve this, we could use refresh of the NCU as a way to tell NWAM to
> attempt
> to reacquire DHCP lease for drivers - a refresh request will do the job, since
> a refresh will do a DHCP release, unplumb and replumb and kick off DHCP again,
> and when this times out, we fall back to wireless. However this whole cycle
> requires user intervention and is time-consuming. This would involve exposing
> refresh actions to libnwam consumers as an nwam_ncu_refresh() function and
> adding a button to the GUI to do this.
An important detail here is that if NWAM invokes DHCP, the dhcpagent daemon
will pick up a previously cached lease if it can't contact the server and if
one is available. This is a hazard when working with DHCP -- it can
unexpectedly "succeed" without actually doing anything useful. In fact, it's
downright harmful, because it can't tell the distinction between "server is
temporarily unavailable, but address is just fine" on one hand, and "interface
itself is completely non-functional" on the other.
Avoiding this failure mode means clearing out the cache first, either by doing
"dhcp release" rather than "dhcp drop" in the first place (hard to arrange in
all cases), or by nuking the /etc/dhcp/${INTF}.dhc file before starting DHCP
again.
--
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.
You are the assignee for the bug.
