NM disconnects Wi-Fi every 2 minutes but iwconfig+dhclient and wicd don't -- ath9k

Thank you for taking the time to report this bug and helping to make Ubuntu better. Unfortunately, we can't fix it without a bit more information.

* Disable Network manager
* Run tail -n0 -f /var/log/syslog > buglog (from terminal)
* Enable Network manager
* Connect to network
* Wait for disconnect
* Stop capturing do Ctrl-C (in terminal)

And upload the file in your /home dir and attached.

Thanks!

I made a log...
Dec 13 14:23:30 szaraeminencja6 wpa_supplicant[1303]: CTRL-EVENT-SCAN-RESULTS
is a problem..
it does not seem to disconnect but instead starts a scan... I´m not sure if it was like that before (I was quite sure I was being disconnected)... however it clearly is now
Each time CTRL-EVENT-SCAN-RESULTS happens the whole network stops for around 1 s (I followed it with ping) and then goes on again...

The work around is still to disable network manger and configure wifi card with iwconfig and dhclient manually...

I believe this bug is now fixed in the latest devlopment release, Lucid, kernel2.6.32 - where this certain driver was being forced to rescan causing the dissconnection. Would it be possible if you could download the alpha1 livecd, set up the driver on the livecd and see if the new kernel fixes the issue.

I did a test on Lucid...
and there is no much change. Scan events are still there and they are stopping/slowing network down although much less then on 9.10 (pinglog from Lucid also attached).

Please see

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/461769

if it's any help

I have this same problem. I need to be able to use network manager b/c I roam a lot and it would be near impossible to manually connect for every single wifi network I try to connect to.

Seems to be still issue in Maverick Meerkat

[Expired for network-manager (Ubuntu) because there has been no activity for 60 days.]

I observe the same behaviour on Maverick Meerkat with kernel 2.6.35-25-generic and
network-manager 0.8.1+git.20100810t184654.ab580f4-0ubuntu2.

The connection using network-manager works for a few seconds and then stops working.
Everything works fine when deactivating network-manager and using wpa_supplicant.

What additional information would be needed to debug this problem ?

Thanks!

Status changed to 'Confirmed' because the bug affects multiple users.

I'm seeing this as well, in ubuntu 11.04. Switching to wicd seems to fix the problem, so it must be an issue specific to network manager. After I restart network manager (/etc/init.d/network-manager restart) it seems to work briefly (a few seconds). Just long enough to connect to the network. Then it disconnects again and the syslog starts filling up again. Amazing that this bug has been around so long! Good thing there's wicd....

deauth by local choice (reason = 3); most likely because the background scans are causing this particular driver to deauth from the AP for scanning.

Is this still the case in 12.04?

I have always blamed the driver for this (and nothing happend from that side), but seeing this might change my mind. It is very frustrating to use ath9k (even the latest version, manually installed) and not to have a reliable internet connection.

It is not every 2 minutes, it is once a day or maybe every minute, depending on the activity. For me the only partial solution seems to be resetting the router, as otherwise I cannot log to the wifi connection (with WPA2 security).

And as a completion: this is very old, at least since Natty (the time I switched to linux). I now use Ubuntu Precise (12.04).

uname -a
Linux florin-Satellite-C650 3.2.0-26-generic #41-Ubuntu SMP Thu Jun 14 17:49:24 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

@florin: Can you please attach a new syslog covering one of the disconnections?

@Thomas Hood (#17)
Please tell exactly the command to run in Terminal, I am a simple user :)

@florin: Hi. The best thing to do is: Reboot, reproduce the malfunction, then open a terminal and run "apport-collect 495981" which will attach a lot of relevant information to this bug report.

@Thomas Hood (#19)

I tried that, but (of course) it does not work as when the bug occurs, I do not have an active internet connection.

Restarted the wifi router and ran again the command. This is the output:

You are not the reporter or subscriber of this problem report, or the report is a duplicate or already closed.

Please create a new report using "apport-bug".

I just subscribed you to this report, so now I hope apport-collect will work. Please reboot, reproduce the malfunction, then open a terminal and run "sudo apport-collect 495981".

ApportVersion: 2.0.1-0ubuntu11
Architecture: amd64
DistroRelease: Ubuntu 12.04
IfupdownConfig:
 auto lo
 iface lo inet loopback
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Release amd64 (20120425)
IpRoute:
 default via 192.168.2.1 dev wlan0 proto static
 169.254.0.0/16 dev wlan0 scope link metric 1000
 192.168.2.0/24 dev wlan0 proto kernel scope link src 192.168.2.102 metric 2
NetworkManager.state:
 [main]
 NetworkingEnabled=true
 WirelessEnabled=true
 WWANEnabled=true
 WimaxEnabled=true
Package: network-manager 0.9.4.0-0ubuntu4.1
PackageArchitecture: amd64
ProcVersionSignature: Ubuntu 3.2.0-27.43-generic 3.2.21
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
Tags: precise
Uname: Linux 3.2.0-27-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups:

nmcli-dev:
 DEVICE TYPE STATE DBUS-PATH
 wlan0 802-11-wireless connected /org/freedesktop/NetworkManager/Devices/1
 eth0 802-3-ethernet unavailable /org/freedesktop/NetworkManager/Devices/0
nmcli-nm:
 RUNNING VERSION STATE NET-ENABLED WIFI-HARDWARE WIFI WWAN-HARDWARE WWAN
 running 0.9.4.0 connected enabled enabled enabled enabled disabled

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

All the above attachments are uploaded after the bug occured and I reset the router to have internet access. Right now, a few minutes later, I have it again. It's not a great computer experience... :(

I can provide with any log and other information one would need to fix this, even remote control to this computer (but secured).

ApportVersion: 2.0.1-0ubuntu11
Architecture: amd64
DistroRelease: Ubuntu 12.04
IfupdownConfig:
 auto lo
 iface lo inet loopback
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Release amd64 (20120425)
IpRoute:
 default via 192.168.2.1 dev wlan0 proto static
 169.254.0.0/16 dev wlan0 scope link metric 1000
 192.168.2.0/24 dev wlan0 proto kernel scope link src 192.168.2.102 metric 2
NetworkManager.state:
 [main]
 NetworkingEnabled=true
 WirelessEnabled=true
 WWANEnabled=true
 WimaxEnabled=true
Package: network-manager 0.9.4.0-0ubuntu4.1
PackageArchitecture: amd64
ProcVersionSignature: Ubuntu 3.2.0-27.43-generic 3.2.21
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
Tags: precise
Uname: Linux 3.2.0-27-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups:

nmcli-dev:
 DEVICE TYPE STATE DBUS-PATH
 wlan0 802-11-wireless connected /org/freedesktop/NetworkManager/Devices/1
 eth0 802-3-ethernet unavailable /org/freedesktop/NetworkManager/Devices/0
nmcli-nm:
 RUNNING VERSION STATE NET-ENABLED WIFI-HARDWARE WIFI WWAN-HARDWARE WWAN
 running 0.9.4.0 connected enabled enabled enabled enabled disabled

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

The last set of attachments are after the second router reset and "sudo apport-collect 495981". Maybe some other way of collecting local data while the bug is active would be better? In this version, I can only run this command if I reset the router and I reconnect on the wifi connection

How this bug can still be going today amazes me. I experience this type of unstability on my ASUS UX31E as well. Wireless card is an Ahteros AR9485.

I apologize if this is the wrong bug.

I do have the same problem. On my machine the Networkmanager links up with a SSID that is the strongest access point (of my neighbors) , but that access point is not configured. Hence, the link dies, takes some time to recover and then the sequence starts again after two minutes. The bug is that Networkmanager should not make a connection to a not configured access point even if that has a better signal.

In debug mode.

The wrong SSID:
 wpa_supplicant[1259]: wlan0: BSS: Expire BSS 13 due to age
 wpa_supplicant[1259]: wlan0: BSS: Remove id 13 BSSID 00:1a:70:47:b0:de SSID 'linksys'

The right SSID:
wpa_supplicant[1259]: wlan0: 0: b0: : : : ssid='Zadel' wpa_ie_len=0 rsn_ie_len=20 caps=0xc11 level=10

With the command iwevent the two minute scan reveals itself.

So, if there is no stronger access point than the configured one, the problem does not surface.

This problems exists up to kernel 3.11.0-3-generic #8-Ubuntu SMP Fri Aug 23 16:49:15 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

I think there is a second bug connected with comment 51. That is that Networkmanager THINKS that an access point is stronger. I checked my comment an other machine. On that machine the wrong SSID 'linksys' was NOT stronger. On my machine it was.
After a rescan the strength of the connection with the wrong SSID dropped on my machine too and my problems went (temporarily) away.

So perhaps this bug is ALSO triggered by an other bug that gives false readings of signal strength and triggers an incomplete connection.

Adding to comment 52. The frequencies of the access point are the same. In this example Channel 11. Is the networkmanager getting confused and makes it a connection to the incorrect access point? All evidence points in this direction. It explains also the wild changing of the signal strength.

A work around is to change the channel of you own access point to an unused channel. If there is one.

wlan0 Scan completed :
          Cell 01 - Address: :::::
                    Channel:11
                    Frequency:2.462 GHz (Channel 11)
                    Quality=70/70 Signal level=10 dBm
                    Encryption key:on
                    ESSID:"Zadel"

          Cell 04 - Address: :::::
                    Channel:11
                    Frequency:2.462 GHz (Channel 11)
                    Quality=67/70 Signal level=-43 dBm
                    Encryption key:off
                    ESSID:"Sitecom

I switched to channel 13. Less problems but if the Internet connection is (shorter) blocked I get:

Or:
sudo iwlist wlan0 scan
wlan0 Interface doesn't support scanning : Device or resource busy

Or:
wlan0 Scan completed :
          Cell 01 - Address: B0:B2:DC:10:33:6C
                    Channel:13
                    Frequency:2.472 GHz (Channel 13)
                    Quality=70/70 Signal level=10 dBm
                    Encryption key:on
                    ESSID:"Zadel"

Cell 06 - Address: 00:0C:F6:42:DD:9A
                    Channel:11
                    Frequency:2.462 GHz (Channel 11)
                    Quality=70/70 Signal level=10 dBm
                    Encryption key:off
                    ESSID:"Sitecom"

The Quality is the same and that should not be possible. So, something very low level is going on. It is beyond my capabilities to debug this.

The problem is driver bound. That is, some drivers never have this problem. At my place an Intel N2230 driver. Other drivers have a lot of problems. That my home Rtl8192se drivers.

So, it is still not clear what causes this bug.

FWIW, still broken on 14.04, the workaround of manually fiddling with the NetworkManager config still helps.

Ah, durn, wrong bug, sorry.

I added a registration to iw:

sudo iw reg set US

This seems to have solved many of the problems that I have been having with random disconnect

The networkmanager connects at my place to BSSID none. None does not exists, so the connection is lost and restarted again, and again and again.

The relevant log is:

<info> (wlan0): roamed from BSSID (none) ((none)) to 7C: (Zadel6)
wpa_supplicant[1076]: wlan0: WPA: Key negotiation completed with 7c: [PTK=CCMP GTK=TKIP]
wlan0: deauthenticated from 7c:(Reason: 15=4WAY_HANDSHAKE_TIMEOUT)
wpa_supplicant[1076]: message repeated 6 times: [ wlan0: WPA: Key negotiation completed with 7c:[PTK=CCMP GTK=TKIP]]
wpa_supplicant[1076]: wlan0: CTRL-EVENT-DISCONNECTED bssid=7c: reason=15
<warn> Connection disconnected (reason 15)

By chance I found the cause. My Cisco router

Firmware Version: epc3928a-E10-5-v302r125572-131030c-ZIG
MAC Address:
Current Time: Sat. Nov 22 11:42:30 2014
Router Mode: IPv4

has two bands 2,4 Ghz and 5Ghz, both with the same BSSID Zadel6.

In this situation the networkmanager is switching from BSSID Zadel6 to none, none to none and none to Zadel6.
If I disable one of the bands or give it a different BSSID, the switching disappears and the connection is not lost anymore.

If it correct to conclude that the networkmanager is getting confused by the two bands on the same BSSID?

Good Morning,

I'm recording in this report that I am also being affected by the bug described above. Remembering , reported 4 years ago .

Thanks.

I'm going to bump this, as the last report was years ago. I'm now experiencing this issue on a Qualcomm Atheros AR938x device I've had for years now. Same issue with constant disconnects as well as a new issue that might involve iommu?

[ 680.361521] ath9k 0000:21:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfdee1544 flags=0x0000]
[ 680.425700] ath9k 0000:21:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfe720fc4 flags=0x0000]
[ 680.425707] ath9k 0000:21:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfe721000 flags=0x0000]
[ 680.489520] ath9k 0000:21:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xff0d5a44 flags=0x0000]
[ 680.650051] wlo1: authenticate with 10:c3:7b:ce:7e:ac
[ 680.665173] wlo1: send auth to 10:c3:7b:ce:7e:ac (try 1/3)
[ 680.666229] wlo1: authenticated
[ 680.669072] wlo1: associate with 10:c3:7b:ce:7e:ac (try 1/3)
[ 680.670356] wlo1: RX AssocResp from 10:c3:7b:ce:7e:ac (capab=0x11 status=0 aid=4)
[ 680.670494] wlo1: associated
[ 1014.692234] wlo1: authenticate with 10:c3:7b:ce:7e:ac
[ 1014.707630] wlo1: send auth to 10:c3:7b:ce:7e:ac (try 1/3)
[ 1014.708714] wlo1: authenticated
[ 1014.711340] wlo1: associate with 10:c3:7b:ce:7e:ac (try 1/3)
[ 1014.712644] wlo1: RX AssocResp from 10:c3:7b:ce:7e:ac (capab=0x11 status=0 aid=4)
[ 1014.712784] wlo1: associated
[ 1295.496298] wlo1: authenticate with 10:c3:7b:ce:7e:ac
[ 1295.511421] wlo1: send auth to 10:c3:7b:ce:7e:ac (try 1/3)
[ 1295.512484] wlo1: authenticated
[ 1295.514554] wlo1: associate with 10:c3:7b:ce:7e:ac (try 1/3)
[ 1295.515851] wlo1: RX AssocResp from 10:c3:7b:ce:7e:ac (capab=0x11 status=0 aid=4)

Same problem on Ubuntu 22.04
corrado@corrado-n3-jj-0103:~$ apt policy network-manager
network-manager:
  Installed: 1.32.12-0ubuntu2
  Candidate: 1.32.12-0ubuntu2
  Version table:
 *** 1.32.12-0ubuntu2 500
        500 http://archive.ubuntu.com/ubuntu jammy/main amd64 Packages
        100 /var/lib/dpkg/status
corrado@corrado-n3-jj-0103:~$

Hi everybody !

I've got exactly the same trouble:

HP-14-TS:~$ apt policy network-manager
network-manager:
  Installé : 1.36.4-2ubuntu1
  Candidat : 1.36.4-2ubuntu1
 Table de version :
 *** 1.36.4-2ubuntu1 500
        500 http://fr.archive.ubuntu.com/ubuntu jammy/main amd64 Packages
        100 /var/lib/dpkg/status
HP-14-TS:~$

Thank you in advance for your help

Regards,
Sab
wlo1 IEEE 802.11 ESSID:
          Mode:Managed Frequency:2.427 GHz Access Point: A4:3E:51:ED:20:B9
          Bit Rate=54 Mb/s Tx-Power=15 dBm
          Retry short limit:7 RTS thr:off Fragment thr:off
          Power Management:off
          Link Quality=59/70 Signal level=-51 dBm
          Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
          Tx excessive retries:0 Invalid misc:135 Missed beacon:0