I still get this bizarre behaviour, though with a slight variation.
I get a connection OK, but about a minute or so later, the interface
gets another lease from a dhclient that NM started, but nevertheless
doesn't regard as a belonging to it (unrecognised PID). NM then *stops*
the interface.
Better wording... why not accept a lease whereever it arrives from?
What seems to be happening is that NM *ignores* the provided lease
because it's mistakenly waiting for a lease from a different PID.
Things then time out.
Easy workaround: just accept the lease.
Fix: track them PIDs correctly.
i never could reproduce this. one idea would be that you have some hooks
in /etc/network/if-*.d/ that run dhclient; or something else that runs
dhclient when a network device is upped.
If you still see this reopen and subscribe me directly.
(btw, sorry for the radar droppage)
** Changed in:
** Summary changed:
- Network-Manager requests DHCP configuration twice, gets confused and aborts
the connection
+ 0.7 Network-Manager requests DHCP configuration twice, gets confused and
aborts the connection
** Description changed:
Binary package hint: network-manager
Here is an
can you reproduce that after a reboot? do you have more than one
NetworkManager process running at that time?
** Changed in: network-manager (Ubuntu)
Status: New = Incomplete
--
0.7 Network-Manager requests DHCP configuration twice, gets confused and aborts
the connection
Now it won't even connect after i hit OK to try again.
Probably a different bug, but I'm losing track of how many different
bugs I'm hitting here so this'll do.
** Attachment added: syslog
http://launchpadlibrarian.net/16625392/fresh%20log
** Changed in: network-manager (Ubuntu)