Hi.

So NetworkManager was giving me lots of debugging output. Having a look
on Google shows it doesn't have log levels, so my only option (apart
from turning it off) was to redirect its log output to a separate file.
It's growing quickly (remind me to set up logrotate!)

After I've removed NetworkManager from /var/log/messages, I'm left with:
Jan 29 23:45:59 orpheus .:
Jan 29 23:46:05 orpheus .:
Jan 29 23:46:11 orpheus .:
Jan 29 23:46:17 orpheus .:
Jan 29 23:46:23 orpheus .:
Jan 29 23:46:29 orpheus .:

and so on every 6 seconds. I think the "." is supposed to be the
program name, and obviously after the ":" comes the message "".

Every message seems to exactly correspond to this NetworkManager
message:
Jan 29 23:48:17 orpheus NetworkManager: <WARN> killswitch_getpower_reply(): Error getting killswitch power: Method "GetPower" with signature "" on interface "org.freedesktop.Hal.Device.KillSwitch" doesn't exist
Jan 29 23:48:17 orpheus NetworkManager: <WARN> killswitch_getpower_reply(): Error getting killswitch power: dellWirelessCtl (/usr/bin/dellWirelessCtl) not available or executable.

Note the reference to /usr/bin/dellWirelessCtl (which really doesn't
exist!) If I symlink /usr/bin/dellWirelessCtl to /bin/false, all the
log messages stop and the last one says:

Jan 29 23:49:35 orpheus NetworkManager: <info> Wireless now disabled by radio killswitch

If I leave the symlink there, NetworkManager won't bring up wireless, so
it's a useless "hack".

I don't really care about any killswitch operation, but I'm interested
in why I'm getting a "." message. NetworkManager bug or
misconfiguration error?

TIA,
--
Iain Buchanan <iaindb at netspace dot net dot au>

I do not know myself and God forbid that I should.
-- Johann Wolfgang von Goethe

On Friday 29 January 2010 16:26:42 Iain Buchanan wrote:
> Hi.
>
> So NetworkManager was giving me lots of debugging output. Having a look
> on Google shows it doesn't have log levels, so my only option (apart
> from turning it off) was to redirect its log output to a separate file.
> It's growing quickly (remind me to set up logrotate!)
>
> After I've removed NetworkManager from /var/log/messages, I'm left with:
> Jan 29 23:45:59 orpheus .:
> Jan 29 23:46:05 orpheus .:
> Jan 29 23:46:11 orpheus .:
> Jan 29 23:46:17 orpheus .:
> Jan 29 23:46:23 orpheus .:
> Jan 29 23:46:29 orpheus .:
>
> and so on every 6 seconds. I think the "." is supposed to be the
> program name, and obviously after the ":" comes the message "".
>
> Every message seems to exactly correspond to this NetworkManager
> message:
> Jan 29 23:48:17 orpheus NetworkManager: <WARN>
> killswitch_getpower_reply(): Error getting killswitch power: Method
> "GetPower" with signature "" on interface
> "org.freedesktop.Hal.Device.KillSwitch" doesn't exist Jan 29 23:48:17
> orpheus NetworkManager: <WARN> killswitch_getpower_reply(): Error getting
> killswitch power: dellWirelessCtl (/usr/bin/dellWirelessCtl) not available
> or executable.
>
> Note the reference to /usr/bin/dellWirelessCtl (which really doesn't
> exist!) If I symlink /usr/bin/dellWirelessCtl to /bin/false, all the
> log messages stop and the last one says:
>
> Jan 29 23:49:35 orpheus NetworkManager: <info> Wireless now disabled by
> radio killswitch
>
> If I leave the symlink there, NetworkManager won't bring up wireless, so
> it's a useless "hack".
>
> I don't really care about any killswitch operation, but I'm interested
> in why I'm getting a "." message. NetworkManager bug or
> misconfiguration error?

Run syslog-ng with the -d switch to enable it's debug output (normally to
messages), or use -dd to get even more debug output.

Beware, this adds up real quick, so don't run it for long like that. The
output may give you more of a clue as to what syslog-ng thinks the incoming
messages are.



--
alan dot mckinnon at gmail dot com

On Fri, 2010-01-29 at 17:29 +0200, Alan McKinnon wrote:
> On Friday 29 January 2010 16:26:42 Iain Buchanan wrote:

> > I don't really care about any killswitch operation, but I'm interested
> > in why I'm getting a "." message. NetworkManager bug or
> > misconfiguration error?
>
> Run syslog-ng with the -d switch to enable it's debug output (normally to
> messages), or use -dd to get even more debug output.
>
> Beware, this adds up real quick, so don't run it for long like that. The
> output may give you more of a clue as to what syslog-ng thinks the incoming
> messages are.

Holy Debug Messages, Batman! Sure does add up real quick.

56,599 messages all with the same timestamp Feb 2 11:13:00; 100% cpu
usage, and 200+Mb before I killed it.

Shirley that's not right?

The 50k of messages all look like this:

Feb 2 11:12:59 orpheus syslog-ng[3739]: Filter rule evaluation begins; filter_rule='f_networkmanager'
Feb 2 11:12:59 orpheus syslog-ng[3739]: Filter node evaluation result; filter_result='not-match'
Feb 2 11:12:59 orpheus syslog-ng[3739]: Filter rule evaluation result; filter_result='not-match', filter_rule='f_networkmanager'

my syslog conf is directing network manager to a separate file:

@version: 3.0
# $Header: /var/cvsroot/gentoo-x86/app-admin/syslog-ng/files/syslog-ng.conf.gentoo.3.0,v 1.1 2009/05/25 20:07:21 mr_bones_ Exp $
#
# Syslog-ng default configuration file for Gentoo Linux

options {
chain_hostnames(no);

# The default action of syslog-ng is to log a STATS line
# to the file every 10 minutes. That's pretty ugly after a while.
# Change it to every 12 hours so you get a nice daily update of
# how many messages syslog-ng missed (0).
stats_freq(43200);
};

source src {
unix-stream("/dev/log" max-connections(256));
internal();
file("/proc/kmsg");
};

destination messages { file("/var/log/messages"); };

# By default messages are logged to tty12...
destination console_all { file("/dev/tty12"); };
# ...if you intend to use /dev/console for programs like xconsole
# you can comment out the destination line above that references /dev/tty12
# and uncomment the line below.
#destination console_all { file("/dev/console"); };

# NetworkManager log to different file
log {
source(src);
filter(f_networkmanager);
destination(df_networkmanager);
flags(final);
};
log { source(src); destination(messages); };
log { source(src); destination(console_all); };

filter f_networkmanager { program("NetworkManager"); };
destination df_networkmanager { file("/var/log/NetworkManager.log"); };

any ideas? thanks,
--
Iain Buchanan <iaindb at netspace dot net dot au>

A synonym is a word you use when you can't spell the word you first
thought of.
-- Burt Bacharach

On Tuesday 02 February 2010 04:06:14 Iain Buchanan wrote:
> On Fri, 2010-01-29 at 17:29 +0200, Alan McKinnon wrote:
> > On Friday 29 January 2010 16:26:42 Iain Buchanan wrote:
> > > I don't really care about any killswitch operation, but I'm interested
> > > in why I'm getting a "." message. NetworkManager bug or
> > > misconfiguration error?
> >
> >
> > Run syslog-ng with the -d switch to enable it's debug output (normally
> > to messages), or use -dd to get even more debug output.
> >
> > Beware, this adds up real quick, so don't run it for long like that. The
> > output may give you more of a clue as to what syslog-ng thinks the
> > incoming messages are.
>
> Holy Debug Messages, Batman! Sure does add up real quick.
>
> 56,599 messages all with the same timestamp Feb 2 11:13:00; 100% cpu
> usage, and 200+Mb before I killed it.
>
> Shirley that's not right?
>
> The 50k of messages all look like this:
>
> Feb 2 11:12:59 orpheus syslog-ng[3739]: Filter rule evaluation begins;
> filter_rule='f_networkmanager' Feb 2 11:12:59 orpheus syslog-ng[3739]:
> Filter node evaluation result; filter_result='not-match' Feb 2 11:12:59
> orpheus syslog-ng[3739]: Filter rule evaluation result;
> filter_result='not-match', filter_rule='f_networkmanager'

That's definitely not right. Even with full debugging enabled no app should
emit that amount of logs.

Seeing as we are dealing with networkmanager with it's long history of being
hard to deal with, I recommend you

a. recognize the truth - that it is a piece of shit
b. use wicd instead, which is decidedly not a piece of shit

:-)


--
alan dot mckinnon at gmail dot com

On Tuesday 02 February 2010 02:06:14 Iain Buchanan wrote:
> On Fri, 2010-01-29 at 17:29 +0200, Alan McKinnon wrote:
> > On Friday 29 January 2010 16:26:42 Iain Buchanan wrote:
> > > I don't really care about any killswitch operation, but I'm interested
> > > in why I'm getting a "." message. NetworkManager bug or
> > > misconfiguration error?
> >
> > Run syslog-ng with the -d switch to enable it's debug output (normally to
> > messages), or use -dd to get even more debug output.
> >
> > Beware, this adds up real quick, so don't run it for long like that. The
> > output may give you more of a clue as to what syslog-ng thinks the
> > incoming messages are.
>
> Holy Debug Messages, Batman! Sure does add up real quick.
>
> 56,599 messages all with the same timestamp Feb 2 11:13:00; 100% cpu
> usage, and 200+Mb before I killed it.
>
> Shirley that's not right?
>
> The 50k of messages all look like this:
>
> Feb 2 11:12:59 orpheus syslog-ng[3739]: Filter rule evaluation begins;
> filter_rule='f_networkmanager' Feb 2 11:12:59 orpheus syslog-ng[3739]:
> Filter node evaluation result; filter_result='not-match' Feb 2 11:12:59
> orpheus syslog-ng[3739]: Filter rule evaluation result;
> filter_result='not-match', filter_rule='f_networkmanager'
>
> my syslog conf is directing network manager to a separate file:
>
> @version: 3.0
> # $Header:
> /var/cvsroot/gentoo-x86/app-admin/syslog-ng/files/syslog-ng.conf.gentoo.3.
> 0,v 1.1 2009/05/25 20:07:21 mr_bones_ Exp $ #
> # Syslog-ng default configuration file for Gentoo Linux
>
> options {
> chain_hostnames(no);
>
> # The default action of syslog-ng is to log a STATS line
> # to the file every 10 minutes. That's pretty ugly after a while.
> # Change it to every 12 hours so you get a nice daily update of
> # how many messages syslog-ng missed (0).
> stats_freq(43200);
> };
>
> source src {
> unix-stream("/dev/log" max-connections(256));
> internal();
> file("/proc/kmsg");
> };
>
> destination messages { file("/var/log/messages"); };
>
> # By default messages are logged to tty12...
> destination console_all { file("/dev/tty12"); };
> # ...if you intend to use /dev/console for programs like xconsole
> # you can comment out the destination line above that references /dev/tty12
> # and uncomment the line below.
> #destination console_all { file("/dev/console"); };
>
> # NetworkManager log to different file
> log {
> source(src);
> filter(f_networkmanager);
> destination(df_networkmanager);
> flags(final);
> };
> log { source(src); destination(messages); };
> log { source(src); destination(console_all); };
>
> filter f_networkmanager { program("NetworkManager"); };

Could it be that "NetworkManager" should be "networkmanager"? Also try it
without " " and see if it fixes it.

> destination df_networkmanager { file("/var/log/NetworkManager.log"); };

Have you already created this file?

> any ideas? thanks,
>

--
Regards,
Mick

On Wed, 2010-02-03 at 00:05 +0200, Alan McKinnon wrote:
> On Tuesday 02 February 2010 04:06:14 Iain Buchanan wrote:

> > The 50k of messages all look like this:

> That's definitely not right. Even with full debugging enabled no app should
> emit that amount of logs.

and yet with debugging disabled, there's no cpu usage, so perhaps
there's just a problem with my syslog-ng rules?

> Seeing as we are dealing with networkmanager with it's long history of being
> hard to deal with, I recommend you
>
> a. recognize the truth - that it is a piece of shit

I appreciate the humour, but so far for me, it's Just Worked(TM). Even
with this log file annoyance, it's still "working".

> b. use wicd instead, which is decidedly not a piece of shit

I had a look at that, but it doesn't do 2 things that I use
NetworkManager for:
1. mobile broadband (essential for on the road)
2. NetworkManager sends dbus messages that evolution uses to toggle
its online / offline state. I was sick of forever waiting for
evolution to time out because I'd gone offline. (Granted, you
may think evolution is another POS, but it does certain things
that no other mail client can do, but that's another story)

I found a post that suggested in fact iwlagn wasn't reloading properly
after a suspend, so I've added "UnloadModules iwlagn"
to /etc/hibernate/common.conf and so far I haven't seen the spurious log
messages (cross my fingers).

> :-)

thanks,
--
Iain Buchanan <iain at pcorp dot com dot au>

This is the tomorrow you worried about yesterday. And now you know why.

On Wed, 03 Feb 2010 08:57:19 +0930, Iain Buchanan wrote:

> > b. use wicd instead, which is decidedly not a piece of shit
>
> I had a look at that, but it doesn't do 2 things that I use
> NetworkManager for:
> 1. mobile broadband (essential for on the road)

True, it's been on todo for a while. It's no longer an issue for me now
as I use a Mi-Fi 3G modem, which connects to the computer via WiFi
instead of having a dongle sticking out the side waiting to be knocked
off.

> 2. NetworkManager sends dbus messages that evolution uses to toggle
> its online / offline state. I was sick of forever waiting for
> evolution to time out because I'd gone offline. (Granted, you
> may think evolution is another POS, but it does certain things
> that no other mail client can do, but that's another story)

Wicd can run any command or script you want before and after going on and
offline.


--
Neil Bothwick

-Come, come, why they couldn't hit an elephant from this dist-

On Wednesday 03 February 2010 01:27:19 Iain Buchanan wrote:
> On Wed, 2010-02-03 at 00:05 +0200, Alan McKinnon wrote:
> > On Tuesday 02 February 2010 04:06:14 Iain Buchanan wrote:
> > > The 50k of messages all look like this:
> >
> > That's definitely not right. Even with full debugging enabled no app
> > should emit that amount of logs.
>
> and yet with debugging disabled, there's no cpu usage, so perhaps
> there's just a problem with my syslog-ng rules?
>
> > Seeing as we are dealing with networkmanager with it's long history of
> > being hard to deal with, I recommend you
> >
> > a. recognize the truth - that it is a piece of shit
>
> I appreciate the humour, but so far for me, it's Just Worked(TM). Even
> with this log file annoyance, it's still "working".

You're the lucky one :-)

nm seems to work OK on the RedHats and SuSEs of this world, I've not seen many
folk get it work smoothly on Gentoo

> > b. use wicd instead, which is decidedly not a piece of shit
>
> I had a look at that, but it doesn't do 2 things that I use
> NetworkManager for:

In that case, you could retain NetworkManager and tweak your syslogger to
discard the logs. You'd have to be specific in your MATCH otherwise you might
toss too many false positives, but we already know you ignore those messages
anyway
.
> 1. mobile broadband (essential for on the road)
> 2. NetworkManager sends dbus messages that evolution uses to toggle
> its online / offline state. I was sick of forever waiting for
> evolution to time out because I'd gone offline. (Granted, you
> may think evolution is another POS, but it does certain things
> that no other mail client can do, but that's another story)
>
> I found a post that suggested in fact iwlagn wasn't reloading properly
> after a suspend, so I've added "UnloadModules iwlagn"
> to /etc/hibernate/common.conf and so far I haven't seen the spurious log
> messages (cross my fingers).

Unfortunately, that sounds all too realistic. I gave up trying to use suspend
some time ago after battling with wirelss and graphics hardware that wouldn't
suspend/resume reliably. But with 4G of RAM here, I find it doesn't take much
longer to power down/cold start than suspend/resume
--
alan dot mckinnon at gmail dot com

On Wed, 2010-02-03 at 09:37 +0000, Neil Bothwick wrote:
> On Wed, 03 Feb 2010 08:57:19 +0930, Iain Buchanan wrote:
>
> > > b. use wicd instead, which is decidedly not a piece of shit
> >
> > I had a look at that, but it doesn't do 2 things that I use
> > NetworkManager for:
> > 1. mobile broadband (essential for on the road)
>
> True, it's been on todo for a while. It's no longer an issue for me now
> as I use a Mi-Fi 3G modem, which connects to the computer via WiFi
> instead of having a dongle sticking out the side waiting to be knocked
> off.

And how do you power it on the road? Much more hungry to have 2x wifi
going than one usb 3G modem (imho) and many netbooks are integrating
them so you won't have the dongle sticking out any more.

> > 2. NetworkManager sends dbus messages that evolution uses to toggle
> > its online / offline state. I was sick of forever waiting for
> > evolution to time out because I'd gone offline. (Granted, you
> > may think evolution is another POS, but it does certain things
> > that no other mail client can do, but that's another story)
>
> Wicd can run any command or script you want before and after going on and
> offline.

you're suggestions on exactly what script to run to tell the current evo
process to go offline immediately is welcome I couldn't figure it
out, but no doubt theres some way I could emulate the dbus message from
NetworkManager...

--
Iain Buchanan <iaindb at netspace dot net dot au>

America: born free and taxed to death.

On Wed, 2010-02-03 at 18:57 +0200, Alan McKinnon wrote:
> On Wednesday 03 February 2010 01:27:19 Iain Buchanan wrote:

> > I appreciate the humour, but so far for me, it's Just Worked(TM). Even
> > with this log file annoyance, it's still "working".
>
> You're the lucky one :-)
>
> nm seems to work OK on the RedHats and SuSEs of this world, I've not seen many
> folk get it work smoothly on Gentoo

Wow, I must remember to buy a lottery ticket on the way home
Seriously, it was just emerge and go!

> > I found a post that suggested in fact iwlagn wasn't reloading properly
> > after a suspend, so I've added "UnloadModules iwlagn"
> > to /etc/hibernate/common.conf and so far I haven't seen the spurious log
> > messages (cross my fingers).
>
> Unfortunately, that sounds all too realistic. I gave up trying to use suspend
> some time ago after battling with wirelss and graphics hardware that wouldn't
> suspend/resume reliably. But with 4G of RAM here, I find it doesn't take much
> longer to power down/cold start than suspend/resume

really? 4G RAM, Core 2 Duo T9500 @ 2.60GHz here, and hibernate is much
faster. Do you have an SSD? Resuming with gnome, compiz, firefox, etc.
already loaded is supremely better than my boot up AND log-in time
otherwise.

--
Iain Buchanan <iaindb at netspace dot net dot au>

"Pets are always a great help in times of stress. And in times of
starvation too, o'course."
-- (Terry Pratchett, Small Gods)