suspend/wireless_before_suspend tries to disconnect an inactive device

[Vadim Rutkovsky]

I've got eth0 (ethernet) and wlan0 (wifi) interfaces with eth0 down and wlan0 up.
During tests suspend/wireless_before_suspend fails with error:

Error: Device 'eth0' (/org/freedesktop/NetworkManager/Devices/1) disconnecting failed: This device is not active ** WARNING **: handle_property_changed: failed to update property 'active-connections' of object type NMClient. Error: Timeout 90 sec
expired.

Chechbox should not try to disconnect already inactive devices

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: checkbox 0.13.3
ProcVersionSignature: Ubuntu 3.2.0-17.27-generic 3.2.6
Uname: Linux 3.2.0-17-generic x86_64
NonfreeKernelModules: wl
ApportVersion: 1.93-0ubuntu2
Architecture: amd64
CasperVersion: 1.304
CheckboxSubmission: c029f392d7891201d16e32d195c2666d
CheckboxSystem: b845c366ea09c60efa3a45c1b5b21525
Date: Tue Feb 28 21:11:46 2012
LiveMediaBuild: Ubuntu 12.04 LTS "Precise Pangolin" - Beta amd64 (20120228.1)
ProcEnviron:
 TERM=xterm
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: checkbox
UpgradeStatus: No upgrade log present (probably fresh install)

[Launchpad Janitor (janitor)]

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

[Brendan Donegan (brendan-donegan)]

It's true that the script shouldn't try and disconnect already inactive devices, but at the same time the whole concept of testing wireless before and after suspend depends on their being a pre-existing wireless connection. You *should* have been prompted to create one before this test ran (by virtue of the networking/wireless_connection running). Can you confirm if that was the case? If not we'll need to fix things so that networking/wireless_connection does run beforehand.

I'm not saying this isn't a bug, but at best it's a matter of just prettifying the error - the test will and should fail anyway.

[Brendan Donegan (brendan-donegan)]

Gah, ignore that comment. As you mentioned you do have wlan0 activated...

[Brendan Donegan (brendan-donegan)]

Having got to the bottom of this issue, this is actually not a bug in Checkbox, but the nmcli tool used to drive the test. This:

Error: Device 'eth0' (/org/freedesktop/NetworkManager/Devices/1) disconnecting failed: This device is not active

is not what is causing the test to fail. This is:

** WARNING **: handle_property_changed: failed to update property 'active-connections' of object type NMClient. Error: Timeout 90 sec
expired.

and it's (mostly) unrelated.

We can still change the test so that it doesn't try and do pointless stuff like deactivating already inactive connections but that isn't the crux of the issue and isn't a considerable problem.

[Launchpad Janitor (janitor)]

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

[Brendan Donegan (brendan-donegan)]

ApportVersion: 1.94.1-0ubuntu2
Architecture: amd64
DistroRelease: Ubuntu 12.04
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120121)
Package: network-manager
PackageArchitecture: amd64
ProcEnviron:
 LANGUAGE=en_GB:en
 TERM=xterm
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 3.2.0-18.29-generic 3.2.9
Tags: precise
Uname: Linux 3.2.0-18-generic x86_64
UpgradeStatus: Upgraded to precise on 2012-02-15 (28 days ago)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo

[Brendan Donegan (brendan-donegan)]

apport information

[Brendan Donegan (brendan-donegan)]

apport information

[Brendan Donegan (brendan-donegan)]

apport information

[Brendan Donegan (brendan-donegan)]

Here is the contents of syslog from when the scenario is run (wifi disconnected and then reconnected using nmcli con up). Network Manager has it's log level set to debug

[Brendan Donegan (brendan-donegan)]

This is working in the latest network-manager package. Thanks!

[Brendan Donegan (brendan-donegan)]

The Checkbox bug tasks can remain open to be taken care of at some point.

[Daniel Manrique (roadmr)]

I'll mark this as Fix Released, since checkbox rev 1214 removed the eth0 down part of the wireless tests, and now explicitly passes the tested interface's name to internet_test.