Bug 749988 - Network gets lost pretty often
Network gets lost pretty often
Status: CLOSED DUPLICATE of bug 785239
Product: Fedora
Classification: Fedora
Component: kernel (Show other bugs)
16
Unspecified Unspecified
unspecified Severity unspecified
: ---
: ---
Assigned To: Kernel Maintainer List
Fedora Extras Quality Assurance
:
Depends On:
Blocks:
  Show dependency treegraph
 
Reported: 2011-10-29 11:03 EDT by hannes
Modified: 2012-06-06 20:26 EDT (History)
9 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed: 2012-02-01 15:16:00 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)
dmesg NetworkManager (37.33 KB, text/plain)
2011-10-30 03:21 EDT, hannes
no flags Details
/var/log/messages (76.09 KB, text/plain)
2011-10-31 03:00 EDT, hannes
no flags Details

  None (edit)
Description hannes 2011-10-29 11:03:14 EDT
Description of problem:
Since the upgrade to f16 the network gets lost pretty often.
Wifi:
lspci -nn | grep Network
04:00.0 Network controller [0280]: Intel Corporation WiFi Link 5100 [8086:4232]


Version-Release number of selected component (if applicable):
rpm -qa NetworkManager\*
NetworkManager-glib-0.9.1.90-5.git20110927.fc16.x86_64
NetworkManager-0.9.1.90-5.git20110927.fc16.x86_64
NetworkManager-pptp-0.9.0-1.fc16.x86_64
NetworkManager-gtk-0.9.1.90-5.git20110927.fc16.x86_64
NetworkManager-vpnc-0.9.0-1.fc16.x86_64
NetworkManager-gnome-0.9.1.90-5.git20110927.fc16.x86_64
NetworkManager-openconnect-0.9.0-1.fc16.x86_64

How reproducible:
Erroneous

  
Actual results:
Need to reconnect from time to time.

Expected results:
Constant connectivity like before in f15.


Additional info:
I found the following warning in dmesg:
[ 2607.008078] NetworkManager[680]: NetworkManager[680]: <warn> could not spawn process '/etc/init.d/nscd condrestart': Failed to execute child process "/etc/init.d/nscd" (No such file or directory)
[ 4283.445797] NetworkManager[680]: <warn> could not spawn process '/etc/init.d/nscd condrestart': Failed to execute child process "/etc/init.d/nscd" (No such file or directory)
[ 4283.445805] NetworkManager[680]: NetworkManager[680]: <warn> could not spawn process '/etc/init.d/nscd condrestart': Failed to execute child process "/etc/init.d/nscd" (No such file or directory)
[ 4307.007206] NetworkManager[680]: <warn> could not spawn process '/etc/init.d/nscd condrestart': Failed to execute child process "/etc/init.d/nscd" (No such file or directory)
[ 4307.007426] NetworkManager[680]: NetworkManager[680]: <warn> could not spawn process '/etc/init.d/nscd condrestart': Failed to execute child process "/etc/init.d/nscd" (No such file or directory)

I could not really determine if it's the cause but anyway it doesn't seem right. nscd is installed:
rpm -qa nscd
nscd-2.14.90-14.x86_64
Comment 1 Adam Williamson 2011-10-29 13:57:23 EDT
well, it just looks like NM tries to restart nscd in certain cases (like when the network goes down). I doubt it's the *cause*. The use of 'condrestart' implies that it's only an 'optional' thing and it failing shouldn't be an issue.

What it should be doing now is 'systemctl try-restart nscd.service' , but I doubt fixing that would fix your network.

Can you attach *all* NM logs - something like 'grep NetworkManager /var/log/messages' ? Thanks.



-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers
Comment 2 hannes 2011-10-30 03:20:57 EDT
Well ok, /var/log/messages is empty and I am not sure if syslogd is running:
ps ax | grep syslog
  361 ?        Ss     0:00 /lib/systemd/systemd-kmsg-syslogd
  372 ?        Ss     0:00 /lib/systemd/systemd-stdout-syslog-bridge
 2082 pts/0    S+     0:00 grep syslog

I attach the output of dmesg grepped for NetworkManager. Hope this helps.
Comment 3 hannes 2011-10-30 03:21:30 EDT
Created attachment 530818 [details]
dmesg NetworkManager
Comment 4 Adam Williamson 2011-10-30 13:42:17 EDT
"Well ok, /var/log/messages is empty and I am not sure if syslogd is running:"

systemctl enable rsyslogd.service

well, from the dmesg, it looks like it stays up for 13 minutes, then:

[  805.320176] NetworkManager[681]: NetworkManager[681]: <info> (wlan0): disconnecting for new activation request.

But I'm not totally sure why that would happen. I'd more expect that message if you changed APs, or something.



-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers
Comment 5 hannes 2011-10-31 02:59:59 EDT
Ok needed to install rsyslog-sysvinit and started it with chkconfig, since your command refused to work. Didn't think about this possibility since I thought it is already systemd, sorry.
Just before dropping the connection I have:
Oct 31 07:58:19 caprica systemd[1]: Failed to read PID file /run/sendmail.pid after start. The service might be broken.

Don't know if this is also related. Will attach the whole /var/log/messages since before the disconnect there were no NetworkManager related issues.
Comment 6 hannes 2011-10-31 03:00:32 EDT
Created attachment 530907 [details]
/var/log/messages
Comment 7 Christoph Wickert 2011-10-31 15:33:07 EDT
I have the same problem with F15, also with Intel wifi

$ lspci | grep Network
00:19.0 Ethernet controller: Intel Corporation 82577LM Gigabit Network Connection (rev 06)
02:00.0 Network controller: Intel Corporation Centrino Ultimate-N 6300 (rev 35)

dmesg:
[51344.561558] wlan0: deauthenticating from bc:05:43:81:77:02 by local choice (reason=3)
[51344.580045] cfg80211: Calling CRDA to update world regulatory domain
[51345.039317] cfg80211: World regulatory domain updated:
[51345.039322] cfg80211:     (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
[51345.039328] cfg80211:     (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[51345.039334] cfg80211:     (2457000 KHz - 2482000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)
[51345.039338] cfg80211:     (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)
[51345.039343] cfg80211:     (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[51345.039347] cfg80211:     (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[51345.039368] cfg80211: Calling CRDA for country: DE
[51345.044088] cfg80211: Regulatory domain changed to country: DE
[51345.044093] cfg80211:     (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
[51345.044099] cfg80211:     (2400000 KHz - 2483500 KHz @ 40000 KHz), (N/A, 2000 mBm)
[51345.044103] cfg80211:     (5150000 KHz - 5250000 KHz @ 40000 KHz), (N/A, 2000 mBm)
[51345.044108] cfg80211:     (5250000 KHz - 5350000 KHz @ 40000 KHz), (N/A, 2000 mBm)
[51345.044112] cfg80211:     (5470000 KHz - 5725000 KHz @ 40000 KHz), (N/A, 2698 mBm)

Not sure if this is really NM, but I wonder what "by local choice (reason=3)" is supposed to mean.
Comment 8 hannes 2011-10-31 17:15:35 EDT
Probably something related to Intel drivers. Don't know which component or maintainer could help on a driver-related issue, or should we take it rather upstream (intel) in this case?
Comment 9 Jirka Klimes 2011-12-20 08:30:09 EST
Yeah, that's probably iwlagn driver issue.

You can try whether disabling N-mode helps:
# rmmod iwlagn
# modprobe iwlagn 11n_disable=1

To debug further follow the steps here:
http://live.gnome.org/NetworkManager/Debugging#wifi
That will enable debug logs from wpa_supplicant

Other debugging tips:
* http://intellinuxwireless.org/?n=fw_error_report
* run 'iw event -f -t' to monitor wireless events
Comment 10 hannes 2012-01-30 15:26:16 EST
Ok, it appears again with the new kernel:
3.2.2-1.fc16.x86_64
I have "disconnects" where for example firefox is not able to load pages but network manager is still connected. Most probably an intel driver issue.
04:00.0 Network controller [0280]: Intel Corporation WiFi Link 5100 [8086:4232]
Comment 11 hannes 2012-01-31 15:08:25 EST
Changed the component to kernel since it's most probably a driver issue. I went back to a previous version of the kernel (3.1.10-2.fc16.x86_64), which doesn't have the problem.
Comment 12 Jirka Klimes 2012-02-01 02:47:22 EST
hannes, could you test this:
https://bugzilla.redhat.com/show_bug.cgi?id=785239#c10
Comment 13 hannes 2012-02-01 15:16:00 EST

*** This bug has been marked as a duplicate of bug 785239 ***

Note You need to log in before you can comment on or make changes to this bug.