Debian Jessie pixbuf errors reloaded

They still happen occasionally on dist-upgrade, but recently I was running into them every week – still after one year.

Fix (other than noted on the shell):

# /usr/lib/x86_64-linux-gnu/gdk-pixbuf-2.0/gdk-pixbuf-query-loaders > /usr/lib/x86_64-linux-gnu/gdk-pixbuf-2.0/2.10.0/loaders.cache

Debian Testing: Kernel 3.12 & nvidia dkms 319.82 breaks xorg nvidia 319.76 xserver driver

A few days ago the kernel 3.12+55 has hit Debian Testing (Jessie) so the linux-image-amd64 pulls the new kernel, and therefore also new nvidia-kernel-amd64 packages shipping 379.82.

Problem – the xorg nvidia driver is still 379.76 which breaks the current xserver with kernel 3.12 and nvidia drivers in current Debian Jessie.

# less /var/log/gdm3/:0.log

(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Stereo disabled by request
(**) NVIDIA(0): Enabling 2D acceleration
NVIDIA: API mismatch: the NVIDIA kernel module has version 319.82,
but this NVIDIA driver component has version 319.76.  Please make
sure that the kernel module and all NVIDIA driver components
have the same version.
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module. Please see the
(EE) NVIDIA(0):     system's kernel log for additional error messages and
(EE) NVIDIA(0):     consult the NVIDIA README for details.
(EE) NVIDIA(0):  *** Aborting ***
(EE) NVIDIA(0): Failing initialization of X screen 0
(EE) Screen(s) found, but none have a usable configuration.
(EE)
Fatal server error:
(EE) no screens found(EE)
(EE)
Please consult the The X.Org Foundation support at http://wiki.x.org
 for help.
(EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
(EE)
(EE) Server terminated with error (1). Closing log file.

Temporary solution – remove the current kernel and step down to 3.11 and nvidia 319.76. But beware – that will remove the linux-image-amd64 and nvidia-kernel-amd64 meta packages pulling the latest kernel/nvidia packages too!

# apt-get remove linux-image-3.12-1-amd64
# tail /var/log/apt/history.log
Start-Date: 2014-01-30  19:13:06
Commandline: apt-get remove linux-image-3.12-1-amd64
Remove: linux-image-3.12-1-amd64:amd64 (3.12.6-2), nvidia-kernel-amd64:amd64 (319.82+3.12+1), linux-image-amd64:amd64 (3.12+55), nvidia-kernel-3.12-1-amd64:amd64 (319.82+1+1+3.12.8-1)
End-Date: 2014-01-30  19:13:23

I’m not going the ‘pin the package and get it from debian sid‘ here, since there are too many dependencies pulled.

Debian Testing: Virtualbox 4.3 breaks Vagrant 1.2

Debian Testing pulled a new Virtualbox version (4.3.2) into my system. Which is apparently incompatible with Vagrant 1.2.2

$ vagrant up

Vagrant has detected that you have a version of VirtualBox installed that is not supported. Please install one of the supported versions listed below to use Vagrant:

4.0, 4.1, 4.2

A debian bug is already open, but the package update is only available in sid, not testing.

Thanks lazyfrosch for the hint about pinning unstable and pulling only newer vagrant updates.

# cat < /etc/apt/sources.list.d/unstable.list
deb [arch=amd64,i386] http://ftp.de.debian.org/debian/ sid main non-free contrib
EOF

# cat < /etc/apt/preferences.d/pinning
Package: *
Pin: release a=sid
Pin-Priority: -100

Package: *
Pin: release a=unstable
Pin-Priority: -100

Package: vagrant
Pin: release a=unstable
Pin-Priority: 991
EOF

# apt-get update
# apt-get install vagrant

Now it works like a charm again ๐Ÿ™‚

vagrant_icingaweb2

get_secrets_cb(): Failed to request VPN secrets using gnome-shell, Network Manager & openvpn

Using the openvpn network manager applet, this strange error occured when trying to connect to my VPN. I also figured that the configuration edit was denied with a strange error on not being able to load the template.

Jan  4 16:57:03 nbmif NetworkManager[2872]:  Starting VPN service 'openvpn'...
Jan  4 16:57:03 nbmif NetworkManager[2872]:  VPN service 'openvpn' started (org.freedesktop.NetworkManager.openvpn), PID 6230
Jan  4 16:57:03 nbmif NetworkManager[2872]:  VPN service 'openvpn' appeared; activating connections
Jan  4 16:57:03 nbmif NetworkManager[2872]:  [1388851023.483964] [nm-vpn-connection.c:1374] get_secrets_cb(): Failed to request VPN secrets #3: (6) No agents were available for this request.
Jan  4 16:57:08 nbmif NetworkManager[2872]:  VPN service 'openvpn' disappeared

The gdm session log unveils that something is missing…

$ tail -f .cache/gdm/session.log

      JS LOG: Error 'VPN plugin at /usr/lib/NetworkManager/nm-openvpn-auth-dialog is not executable' while processing VPN keyfile '/etc/NetworkManager/VPN/nm-openvpn-service.name'

Looking at the openvpn service configuration itsself, it’s configured correct:

# vim /etc/NetworkManager/VPN/nm-openvpn-service.name

[GNOME]
auth-dialog=/usr/lib/NetworkManager/nm-openvpn-auth-dialog
properties=/usr/lib/NetworkManager/libnm-openvpn-properties
supports-external-ui-mode=true

But the file itsself is missing…

# ls -la /usr/lib/NetworkManager/nm-openvpn-auth-dialog
ls: Zugriff auf /usr/lib/NetworkManager/nm-openvpn-auth-dialog nicht mรถglich: Datei oder Verzeichnis nicht gefunden

Re-install the involved packages.

# apt-get install --reinstall network-manager-openvpn network-manager-openvpn-gnome

# ls -la /usr/lib/NetworkManager/nm-openvpn-auth-dialog
-rwxr-xr-x 1 root root 31K Sep 13 19:54 /usr/lib/NetworkManager/nm-openvpn-auth-dialog*

Now check the broken symlink for the gnome shell and manually fix it.

# ls -la /usr/lib/gnome-shell/nm-openvpn*
#
# ln -s /usr/lib/NetworkManager/nm-openvpn-auth-dialog /usr/lib/gnome-shell/
# ls -la /usr/lib/gnome-shell/nm-openvpn*
lrwxrwxrwx 1 root root 46 Jan  4 17:10 /usr/lib/gnome-shell/nm-openvpn-auth-dialog -> /usr/lib/NetworkManager/nm-openvpn-auth-dialog*

But still, it remains broken.

$ tail -f .cache/gdm/session.log

      JS LOG: Invalid VPN service type (cannot find authentication binary)

It’s necessary to not only restart the network manager, but also dbus in order to apply the changes. And while at it, save your current work, and reboot. Killing dbus while running a window manager isn’t much fun ๐Ÿ˜‰

# service network-manager restart
# service dbus restart
# reboot

Kudos to this forum post out of many many bug reports and hints. It seems that package upgrade broke/changed the location of the auth dialog file which then broke everything, as usual.

gnome shell flash fullscreen freeze workaround

There’s a bug in the current gnome-shell 3.8.x+ with fullscreen flash videos presented in chromium and iceweasel. While youtube seems to have a workaround for the missing window focus (but only in chromium), other flash players just open the video into an invisible layer, the browser video freezes, and you’ll only hear the sound.

Works in fullscreen – does not work. The bug is described here, but seems a bit dead.

Debian Testing, d-u 2013-12-29
xorg 1.14.5-1
gnome-shell 3.8.4
chrome flash player 11.2 r202

Workaround – use a tool named devilspie and set the focus for such fullscreen events automatically as described here.

# apt-get install devilspie
$ mkdir ~/.devilspie
$ cat <<EOF > ~/.devilspie/flash_fullscreen.ds
(if
(or
    (is (application_name) "plugin-container")
    (and
        (contains application_name) "chromium-browser")
        (contains application_name) "flash-plugin")
    )
)
(begin
(focus)
)
)
EOF

Now make sure that devilspie is started automatically in gnome 3 – either using ‘ALT + F2’ and ‘gnome-session-properties’ and adding a new entry like shown below

autostart_devilspie

or add a new entry manually:

$ cat <<EOF > ~/.config/autostart/devilspie.desktop
[Desktop Entry]
Type=Application
Exec=/usr/bin/devilspie
Hidden=false
NoDisplay=false
X-GNOME-Autostart-enabled=true
Name[de_DE]=devilspie
Name=devilspie
Comment[de_DE]=flash fullscreen fix
Comment=flash fullscreen fix
EOF

Logout/login again, or run devilspie in background using ‘/usr/bin/devilspie&’ – works again ๐Ÿ™‚