GNOME Bugzilla – Bug 341320
nm-applet vanished from tray
Last modified: 2007-09-17 17:05:21 UTC
Please describe the problem: From: https://launchpad.net/distros/ubuntu/+source/network-manager/+bug/43379 The Actors: Fire -- the stalwart DHCP server running Debian Stable (Sarge) Dib -- the noble laptop and DHCP client, possessing a heart of gold and an instance of nm-applet. Dib runs the latest Dapper (May 5th 2006) running nm-applet version 0.6.2-0ubuntu5. Christian -- the mad scientist The Scene: A dark and stormy night. Fire has already given Dib a dynamic IP addressed based on his wifi MAC address. Christian wants to make Dib's IP static when Dib is connected via wifi. What Happened: Christian, sitting at his laptop, Dib, starts to edit the Fire's DHCPd configuration remotely via SSH. He sets Dib's wifi MAC address to return the same IP as the wired MAC does. Christian restarts the DHCP server. The next time Dib asks the DHCP server what it's IP address is, it will change from .189 to .16 A few minutes later, Dib sends the DHCP request. All the SSH connections freeze (as expected) and the nm-applet shows its reconnecting to the network. As nm-applet gets to the end of its reconnect cycle, when it should have shown the wifi bars again, it vanishes. Christian confirms that nm-applet is still running. Christian tries to restart nm-applet, but to no avail... Christian checks the network, it is correctly configured, but nm-applet has lost its mind. A complete reboot is required to make nm-applet work again. What was Expected: It should have "Just Worked". Conclusion: nm-applet or network-manager didn't deal very well with the same network connection handing out a new IP address, even though this is acceptable DHCP behavior. Also, Christian gets into strange moods at time and submits bug reports with a bizarre sense of humor. But at least the reports are proof-read. Steps to reproduce: Actual results: Expected results: Does this happen every time? Other information:
This is probably the daemon exiting/crashing/breaking. nm-applet is supposed to vanish from the tray in that case. If NM is crashing, that is a legit bug, and should be filed with appropriate back trace.
The daemon is crashing on my machine, causing this problem. From the system log on my machine when this happens: Sep 17 09:33:02 2360lnx NetworkManager: <WARN> nm_signal_handler(): Caught signal 11. Generating backtrace... Sep 17 09:33:03 2360lnx NetworkManager: ******************* START ********** ************************ Sep 17 09:33:04 2360lnx NetworkManager: Using host libthread_db library "/li b/tls/i686/cmov/libthread_db.so.1". Sep 17 09:33:05 2360lnx NetworkManager: [Thread debugging using libthread_db enabled] Sep 17 09:33:05 2360lnx NetworkManager: [New Thread -1212422480 (LWP 4385)] Sep 17 09:33:05 2360lnx NetworkManager: [New Thread -1229210736 (LWP 5567)] Sep 17 09:33:05 2360lnx NetworkManager: [New Thread -1220818032 (LWP 4507)] Sep 17 09:33:05 2360lnx NetworkManager: [New Thread -1212425328 (LWP 4501)] Sep 17 09:33:05 2360lnx NetworkManager: 0xffffe410 in __kernel_vsyscall () Sep 17 09:33:05 2360lnx NetworkManager: ******************* END ************ ********************** Though that doesn't look like a useful backtrace. How may one be obtained?