Primus: fatal: failure contacting bumblebee daemon [SOLVED]

Question

I just did a pacman -Syu which broke my system as usual. This time I get the above error message when trying to use primusrun:
primus: fatal: failure contacting bumblebee daemon
Trying to run glxshperes gives a slightly different message:
$ primusrun glxspheres
primus: fatal: Bumblebee daemon reported: error: Could not load GPU driver
and via optirun, just for the heck of it:
$ optirun glxgears
[ 2702.399865] [ERROR]Cannot access secondary GPU - error: Could not load GPU driver
[ 2702.399894] [ERROR]Aborting because fallback start is disabled.
<<SOLUTION: Replace all the *-bumblebee packages by * packages, this means enabling multiverse repo and drawing all of it from official repos instead of anything from AUR. Examples: nvidia instead of nvidia-bumblebee, primus instead of primus-git, nvidia-utils instead of nvidia-utils-bumblebee, lib32-nvidia-utils instead of lib32-nvidia-utils-bumblebee. Some time ago the 'official' instructions were exactly the other way round, but it seems this was now completely reverted.>>
I noticed:
1) primus in official repos (I was using git so far) causes a timeout from about 40 mirrors until it finds one where the download works. I hope that's no bad sign of some sort?
2) bumblebeed update created new bumblebee.conf.pacnew and xorg.conf.nvidia.pacnew
3) off-topic: When I clicked 'logout' in XFCE's applications menu, usually I get a box where I can pick whether I wanna logout/restart/shutdown etc, but this time it just straightly logged me out (makes sense I guess).
4) after rebooting my system I saw I think two systemd error messages pop up for a split second, that usually did not occur.
5) Xorg.8.log had the line "[633217.731] (WW) NVIDIA: This server has an unsupported input driver ABI version (have 19.1, need < 19.0).  The driver will continue to load, but may behave strangely.", is this responsible maybe? Version conflict between X and nvidia?
I tried:
I restarted after the pacman -Syu update.
I tried primusrun-git from AUR and primusrun from <community> but both exit with the error message.
I tried the new conf files (copied the pacnew files to the normal filenames) but they don't work either.
I checked:
Bumbleed is running, systemctl shows
bumblebeed.service loaded active running Bumblebee C Daemon
and lspci | grep VGA shows the usual
00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)
01:00.0 VGA compatible controller: NVIDIA Corporation Device 0fd4 (rev a1)
(After I try to start something with primusrun, the indicator light switches on and remains on, I can turn it off again with the usual
echo OFF > /proc/acpi/bbswitch
dmesg has these:
[ 94.057043] nvidia: disagrees about version of symbol pv_mmu_ops
[ 94.057047] nvidia: Unknown symbol pv_mmu_ops (err -22)
echoing 'OFF' to bbswitch seems fine though (it worked, as I mentioned)
[ 16.326834] bbswitch: version 0.6
[ 16.326840] bbswitch: Found integrated VGA device 0000:00:02.0: \_SB_.PCI0.GFX0
[ 16.326843] bbswitch: Found discrete VGA device 0000:01:00.0: \_SB_.PCI0.PEG0.PEGP
[ 16.326903] bbswitch: detected an Optimus _DSM function
[ 16.326908] bbswitch: Succesfully loaded. Discrete card 0000:01:00.0 is on
[ 16.327995] bbswitch: disabling discrete graphics
Here is /var/log/Xorg.8.log (trimmed):
X.Org X Server 1.14.1
[633217.228] X Protocol Version 11, Revision 0
[633217.228] Build Operating System: Linux 3.8.7-1-ARCH x86_64
[633217.229] (++) Using config file: "/etc/bumblebee/xorg.conf.nvidia"
[633217.229] (==) Using config directory: "/etc/X11/xorg.conf.d"
[633217.293] (++) ModulePath set to "/usr/lib/nvidia-bumblebee/xorg/,/usr/lib/xorg/modules"
[633217.295] Initializing built-in extension XVideo
[633217.295] Initializing built-in extension XVideo-MotionCompensation
[633217.295] Initializing built-in extension XFree86-VidModeExtension
[633217.295] Initializing built-in extension XFree86-DGA
[633217.295] Initializing built-in extension XFree86-DRI
[633217.295] Initializing built-in extension DRI2
[633217.295] (II) LoadModule: "glx"
[633217.308] (II) Loading /usr/lib/nvidia-bumblebee/xorg/modules/extensions/libglx.so
[633217.680] (II) Module glx: vendor="NVIDIA Corporation"
[633217.680] compiled for 4.0.2, module version = 1.0.0
[633217.681] Module class: X.Org Server Extension
[633217.681] (II) NVIDIA GLX Module 313.26 Wed Feb 27 13:10:40 PST 2013
[633217.681] Loading extension GLX
[633217.681] (II) LoadModule: "nvidia"
[633217.705] (II) Loading /usr/lib/xorg/modules/drivers/nvidia_drv.so
[633217.731] (II) Module nvidia: vendor="NVIDIA Corporation"
[633217.731] compiled for 4.0.2, module version = 1.0.0
[633217.731] Module class: X.Org Video Driver
[633217.731] (WW) NVIDIA: This server has an unsupported input driver ABI version (have 19.1, need < 19.0). The driver will continue to load, but may behave strangely.
[633217.732] (II) NVIDIA dlloader X Driver 313.26 Wed Feb 27 12:52:26 PST 2013
[633217.732] (II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
[633217.732] (--) using VT number 7
[633217.733] (II) NVIDIA(0): Creating default Display subsection in Screen section
"Default Screen Section" for depth/fbbpp 24/32
[633217.733] (==) NVIDIA(0): Depth 24, (==) framebuffer bpp 32
[633217.733] (==) NVIDIA(0): RGB weight 888
[633217.733] (==) NVIDIA(0): Default visual is TrueColor
[633217.733] (==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
[633217.733] (**) NVIDIA(0): Option "NoLogo" "true"
[633217.733] (**) NVIDIA(0): Option "UseEDID" "false"
[633217.733] (**) NVIDIA(0): Option "ConnectedMonitor" "CRT-0"
[633217.734] (**) NVIDIA(0): Enabling 2D acceleration
[633217.734] (**) NVIDIA(0): ConnectedMonitor string: "CRT-0"
[633217.734] (**) NVIDIA(0): Ignoring EDIDs
[633218.796] (II) NVIDIA(0): Implicitly enabling NoScanout
[633218.796] (WW) NVIDIA(0): Failed to enable display hotplug notification
[633218.799] (II) NVIDIA(0): NVIDIA GPU GeForce GTX 660M (GK107) at PCI:1:0:0 (GPU-0)
[633218.799] (--) NVIDIA(0): Memory: 2097152 kBytes
[633218.799] (--) NVIDIA(0): VideoBIOS: 80.07.27.00.05
[633218.799] (II) NVIDIA(0): Detected PCI Express Link width: 16X
[633218.799] (--) NVIDIA(0): Valid display device(s) on GeForce GTX 660M at PCI:1:0:0
[633218.799] (--) NVIDIA(0): none
[633218.799] (II) NVIDIA(0): Validated MetaModes:
[633218.799] (II) NVIDIA(0): "nvidia-auto-select"
[633218.799] (II) NVIDIA(0): Virtual screen size determined to be 640 x 480
[633218.799] (WW) NVIDIA(0): Unable to get display device for DPI computation.
[633218.799] (==) NVIDIA(0): DPI set to (75, 75); computed from built-in default
[633218.799] (--) Depth 24 pixmap format is 32 bpp
[633218.799] (II) NVIDIA: Using 3072.00 MB of virtual memory for indirect memory
[633218.799] (II) NVIDIA: access.
[633218.805] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon
[633218.805] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X
[633218.805] (II) NVIDIA(0): configuration option may not be set correctly. When the
[633218.805] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will
[633218.805] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For
[633218.805] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and
[633218.805] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X
[633218.805] (II) NVIDIA(0): Config Options in the README.
[633218.806] (II) NVIDIA(0): Setting mode "nvidia-auto-select"
[633218.857] Loading extension NV-GLX
[633218.865] (==) NVIDIA(0): Disabling shared memory pixmaps
[633218.865] (==) NVIDIA(0): Backing store disabled
[633218.865] (==) NVIDIA(0): Silken mouse enabled
[633218.866] (==) NVIDIA(0): DPMS enabled
[633218.866] Loading extension NV-CONTROL
[633218.866] (II) Loading sub module "dri2"
[633218.866] (II) LoadModule: "dri2"
[633218.866] (II) Module "dri2" already built-in
[633218.866] (II) NVIDIA(0): [DRI2] Setup complete
[633218.866] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia
[633218.866] (--) RandR disabled
[633218.879] (II) Initializing extension GLX
Any ideas?
PS: In this light I was wondering whether the https://wiki.archlinux.org/index.php/Bumblebee article is still up to date or might need reworking of some sort (I don't have any actual ideas to base this on though, was just a thought)
PPS: https://bbs.archlinux.org/viewtopic.php?pid=1264430
Last edited by Jindur (2013-04-27 18:10:21)

Answer 1

iambig wrote:
Hi,
i haved some trouble too, from what i understood
nvidia-bumblebee
is deprecated, you should install
nvidia
otherwise i am not still able to use optirun/primus.
Hm, so far it was emphasized NOT to install nvidia, because that would actually break it, but to use nvidia-bumblebee instead.
Do you maybe have a link to information saying that it is now exactly the other way round?
Has 'nvidia' been confirmed to work with bumblebee by anyone? Does it work for you?
Edit: I just saw in my Xorg.8.log:
[633217.731] (WW) NVIDIA: This server has an unsupported input driver ABI version (have 19.1, need < 19.0). The driver will continue to load, but may behave strangely.
Maybe this has to do with it? It seems there might be a version conflict between X and nvidia drivers?
If anyone can confirm that the basic 'nvidia' packages are indeed replacing 'nvidia-bumblebee' now, instead of the other way round, I'll give them a try
Last edited by Jindur (2013-04-26 12:30:37)