nm-applet crashes if there is no internet: Caught signal 15

open a terminal and kill all

#> sudo killall nm-applet

then start it from command

#> nm-applet

when is crashes please post the output here

Here you are. :)

** Message: applet now removed from the notification area
** Message: applet now embedded in the notification area
** (nm-applet:3867): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:3867): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:3867): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:3867): DEBUG: old state indicates that this was not a disconnect 0
** Message: Caught signal 15, shutting down...

please post the contents of /etc/network/interfaces

if we can't get this working for you then you can uninstall nm-applet and install wicd but let's figure out your problem first, no one ever learned anything from using another package :)

also post the contents of /etc/NetworkManager/nm-system-settings.conf please

//For the network interfaces:

auto lo
iface lo inet loopback

//For the sys-settings:

[main]
plugins=ifupdown,keyfile
[main]
plugins=ifupdown,keyfile

[ifupdown]
managed=true

// Nothing against wicd, but nm is to use better than wicd.. ;)
[ifupdown]
managed=true

oh.. thats wrong..
sys-settings:

[main]
plugins=ifupdown,keyfile

[ifupdown]
managed=true

have you tried uninstall and reinstalling from source?

also, please post lspci

and have you connected via eathernet and updated all packages?

*ethernet

I've purged nm. I was connected via wicd which I've downloaded before.
Nothing has changed.

lspci:
00:00.0 Host bridge: nVidia Corporation MCP79 Host Bridge (rev b1)
00:00.1 RAM memory: nVidia Corporation MCP79 Memory Controller (rev b1)
00:03.0 ISA bridge: nVidia Corporation MCP79 LPC Bridge (rev b3)
00:03.1 RAM memory: nVidia Corporation MCP79 Memory Controller (rev b1)
00:03.2 SMBus: nVidia Corporation MCP79 SMBus (rev b1)
00:03.3 RAM memory: nVidia Corporation MCP79 Memory Controller (rev b1)
00:03.5 Co-processor: nVidia Corporation MCP79 Co-processor (rev b1)
00:04.0 USB Controller: nVidia Corporation MCP79 OHCI USB 1.1 Controller (rev b1)
00:04.1 USB Controller: nVidia Corporation MCP79 EHCI USB 2.0 Controller (rev b1)
00:08.0 Audio device: nVidia Corporation MCP79 High Definition Audio (rev b1)
00:09.0 PCI bridge: nVidia Corporation MCP79 PCI Bridge (rev b1)
00:0b.0 SATA controller: nVidia Corporation MCP79 AHCI Controller (rev b1)
00:0c.0 PCI bridge: nVidia Corporation MCP79 PCI Express Bridge (rev b1)
00:10.0 PCI bridge: nVidia Corporation MCP79 PCI Express Bridge (rev b1)
00:15.0 PCI bridge: nVidia Corporation MCP79 PCI Express Bridge (rev b1)
00:16.0 PCI bridge: nVidia Corporation MCP79 PCI Express Bridge (rev b1)
03:00.0 VGA compatible controller: nVidia Corporation C79 [GeForce G102M] (rev b1)
04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01)
05:00.0 Network controller: Atheros Communications Inc. AR9285 Wireless Network Adapter (PCI-Express) (rev 01)

Maybe you can try downgrading nm-applet and see if that works at least then it will tell us if it is the applet or a program/library it depends on

When I installed nm, I've found that the program has the package number "2". Now is packet 3 installed, but it has not changed.
Should I downgrade to package 1?

Meiko, please run apport-collect 664763, and also attach .xsession-errors, so we can see where things crash. I don't think downgrading the version will do much.

What you can try however, is to kill nm-applet (in console, killall nm-applet) and load a guest session (from the session indicator) -- then start nm-applet if it didn't get started. Any custom configuration you have in your home directory would not impact nm-applet in this case.

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

I have startet the nm-applet in a guest session. I have killed and startet it again via bash. After a little while the same error: ** Message: Caught signal 15, shutting down...
Why it only happens if I am NOT connected to the internet?
Mathieu, how can I show you the last .xsession errors?

Is it doing this after coming back from suspend?

Sometimes he does. But often it is started. If I am connected to the internet, it runs. If not, it crashes after a while.. I think that is the reason why it crashes after coming back from suspend. Sometimes after a few minutes, sometimes at once.

Meiko,

Just add the hidden file .xsession-errors to this bug. Hopefully there will be some more data there.

(polkit-gnome-authentication-agent-1:23489): GLib-GObject-WARNING **: cannot register existing type `_PolkitError'

(polkit-gnome-authentication-agent-1:23489): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed
** Message: killswitch 3 is KILLSWITCH_STATE_UNBLOCKED
** Message: killswitches state KILLSWITCH_STATE_UNBLOCKED
** Message: applet now removed from the notification area
26/10/2010 21:19:59 WARNING: Width (1366) is not a multiple of 4. VncViewer has problems with that.
/usr/share/blueproximity/proximity.py:186: GtkWarning: GtkSpinButton: setting an adjustment with non-zero page size is deprecated
  self.wTree = gtk.glade.XML(self.gladefile)
26/10/2010 21:20:01 Autoprobing TCP port in (all) network interface
26/10/2010 21:20:01 Listening IPv6://[::]:5900
26/10/2010 21:20:01 Listening IPv4://0.0.0.0:5900
26/10/2010 21:20:01 Autoprobing selected port 5900
26/10/2010 21:20:01 Advertising security type: 'TLS' (18)
26/10/2010 21:20:01 Advertising authentication type: 'No Authentication' (1)
26/10/2010 21:20:01 Advertising security type: 'No Authentication' (1)
** (nm-applet:23487): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:23487): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:23487): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:23487): DEBUG: old state indicates that this was not a disconnect 0
** Message: Caught signal 15, shutting down...
Initializing nautilus-gdu extension

(nautilus:23479): GConf-CRITICAL **: gconf_value_free: assertion `value != NULL' failed

Arbeitsflächen-Datei network-manager-gnome.desktop konnte nicht für Panel-Starter geöffnet werden
Starting gtk-window-decorator
I/O warning : failed to load external entity "/home/meiko/.compiz/session/1019da037e7ef3b0b812881207986187700000234010032"

(nautilus:23479): GConf-CRITICAL **: gconf_value_free: assertion `value != NULL' failed

"Arbeitsflächen-Datei network-manager-gnome.desktop konnte nicht für Panel-Starter geöffnet werden"
-- It means "The desktop data "network-manager-gnome.desktop" couldn't open for the panel starter."

okay so from what I've learned so far is

cannot register existing type `_PolkitError'

might be because your on a 64 bit platform?

Can you try running the live cd of 32 and 64 and see if you can recreate the problem?

also was this a fresh install or an upgrade? and how did you do this upgrade if it was

and I doubt this will work but try to kill the nm-applet and restart it using sudo and you can also try gksudo

I don't have a 64 bit platform.
I've installed ubuntu via cd, so it's a fresh install.
I wonder: If I run nm-applet as root, it seems to run stable.. :)

hmm, i wonder if you wrote a script to open nm-applet as root on startup (entering your pass for you so it works) and then run that script to start nm-applet would work...

the fix is to run nm-applet as sudo

thank you! :)

I'm sorry, is it really the "fix" for that? nm-applet never needed root to run, why would it need now? And if indeed needs root (which find highly unlikely), then I'll be forced to type my password everytime I logon just to have the applet show up? Or worse, add an sudoers exception just for that case? That doesn't sound like a fix, sounds like a kludgy workaround.

In my case, running nm-applet as root does not help.

Yep, I tried that "fix" too, and running as sudo in fact didn't work for me either. So, the "fix", apart from being lame, doesn't actually work. Will someone please look into this more carefully? I'm already resigned to the fact that, if this is ever fixed, it'll probably not land in time for Natty Narwhal, but it'll unexcusable if such a glaring and usability-impairing bug slips past Oneiric Ocelot too.

The bug is not fixed.

Setting back to Confirmed since "Opinion" actually means it's closed as being somebody's opinion rather than a bug...

Confirming on the basis that there are multiple people who see this happening, although the exact cause is still unknown.

What would be helpful, is if someone who sees this happen frequently could install debug packages for network-manager-gnome (add the line "deb http://ddebs.ubuntu.com/ maverick main" to /etc/apt/sources.list if running in 10.10, then install network-manager-gnome-dbgsym libnm-glib2-dbgsym libnm-util1-dbgsym libglib2.0-0-dbgsym), and run nm-applet in a debugger:

gdb --args nm-applet

When the debugger starts, type "run" to have nm-applet be started. Once it crashes, type "bt full" then attach the whole output to this bug report.

Added the repository as specified, but on running "apt-get install network-manager-gnome-dbgsym libnm-glib2-dbgsym libnm-util1-dbgsym libglib2.0-0-dbgsym", it can't find the package libglib2.0-0-dbgsym . I installed without it anyway and tried to get a backtrace without it hoping there's something useful, instead of bugging you guys about the missing package and not delivering any bt. Here it is:

(gdb) run
Starting program: /usr/bin/nm-applet
[Thread debugging using libthread_db enabled]
[New Thread 0xb7ddbb70 (LWP 7787)]
** Message: applet now removed from the notification area
** Message: applet now embedded in the notification area
** (nm-applet:7784): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:7784): DEBUG: old state indicates that this was not a disconnect 0
** (nm-applet:7784): DEBUG: old state indicates that this was not a disconnect 2
** (nm-applet:7784): DEBUG: foo_client_state_changed_cb
** (nm-applet:7784): DEBUG: foo_client_state_changed_cb
** (nm-applet:7784): DEBUG: foo_client_state_changed_cb
** (nm-applet:7784): DEBUG: going for offline with icon: notification-network-wireless-disconnected
** (nm-applet:7784): DEBUG: going for offline with icon: notification-network-wireless-disconnected

Program received signal SIGTERM, Terminated.
0x0012e416 in __kernel_vsyscall ()
(gdb) bt full
#0 0x0012e416 in __kernel_vsyscall ()
No symbol table info available.
#1 0x00d43df6 in poll () from /lib/libc.so.6
No symbol table info available.
#2 0x00c01a1b in g_poll () from /lib/libglib-2.0.so.0
No symbol table info available.
#3 0x00bf443c in ?? () from /lib/libglib-2.0.so.0
No symbol table info available.
#4 0x00bf4ba7 in g_main_loop_run () from /lib/libglib-2.0.so.0
No symbol table info available.
#5 0x08059567 in main (argc=1, argv=0xbffff3d4) at main.c:101
        applet = <value optimized out>
        i = <value optimized out>

The 5 log lines before SIGTERM is me disabling the wifi connection in order to have no internet connection, so nm-applet would eventually crash as expected. I see libglib-2.0 is mentioned in the bt, and since I don't have the version of the lib with debug symbols, this bt may not be too useful.

I'm no C programmer, but my impression from all this and earlier debug attempts and peering into nm-applet source code, is that nm-applet is not really crashing - something else is actually killing it. Maybe network-manager itself?

For the libglib debug package, try libglib2.0-0-dbg.

This indeed looks like nm-applet is getting killed, most likely by the kernel's oomkiller whose job is to maintain enough memory on the system. It doesn't mean though that it's necessarily nm-applet leaking memory.

In case it's actually nm-applet leaking memory, it could be a good idea to try to run nm-applet in Valgrind to see how it uses memory in this case. Please see https://wiki.ubuntu.com/Valgrind for instructions on how to do this.

Ok, that went alright now. I installed libc6-dbg too for more completeness. The backtrace is attached.

I'll have a look on Valgrind next, will get back to you as soon as I have more information.

Thanks. Unfortunately, if it's the oomkiller stopping NM-applet, there will not be anything useful in the backtrace. This somewhat goes with what we see here, the applet just receives SIGTERM for an unknown reason.

Keeping this bug open as Incomplete until we can have a valgrind log.

Ok, sorry it took so long to get back. I followed the instructions on the wiki and got nm-applet running from valgrind. The thing is, and that's because it took me so long to get back, nm-applet never crashed when run from valgrind.

I realized that, while running from valgrind, nm-applet run / responded much slower than normal, which is normal according to the wiki.

So I decided to leave it running as long as possible, as I figured that maybe since nm-applet was running slower, that maybe I wasn't allowing it to run long enough to crash. So I left it running for more than 8 hours, without internet connection. It didn't crash even then. I stopped it (CTRL+C), and now I'm attaching the valgrind.log for this 8 hours run - who knows, even though it didn't crash, it might still be useful, though I couldn't spot anything suspicious at a quick glance.

I tested nm-applet again without valgrind and internet connection disabled, and sure enough, it crashed as usual after a couple of minutes.

I'm not sure what to do next... maybe play with the command-line switches on valgrind? Hope you guys have some idea on what else can be done to trace this further.

Regards!

oh, s/crashed/get killed/ , btw... :-P

Thanks. That's pretty good, even with quite a few symbols missing. It's at least something to start with. Because of how valgrind works it probably stopped it from crashing (or it would have crashed, but much later). Setting this to Confirmed/Medium, but so far with no idea how to solve the problem.

I suspect this issue cropped up after upgrading to libc 2.12.1-0ubuntu10.2 which is a dependency of nm. I have been unable to downgrade as synaptic would remove other several other packages and break my installation. Is it possible to revert to libc 2.12.1-0ubuntu6(maverick) through some other means?

What's going on with this?
I'm affected and I think that because nm is crashing I'm getting 2 logouts every day! It is very annoying...

I'm using libc-bin 2.13-0ubuntu13 and networkmanager 0.8.4 on ubuntu 11.04, kernel 2.6.38-10.

This bug affects all the people I know who use Ubuntu 11.10 and also me, because when Ubuntu is going to shutdown it takes 15 more seconds than the previous versions: 400% slower (20 seconds VS 5 seconds)! Isn't the importance High?

My nm-applet crashes sometimes. I run nm-applet in terminal. Check output.

I find this kinda important instead of medium... Is there already improvement visible?

Melroy, please file a new bug report for your problem. Make sure you use the command: "ubuntu-bug network-manager-applet"