0

Trying to install bumblebee on Centos 7.2. My system is an Asus laptop with integrated Intel graphics as well as a Geforce 960M, with no BIOS option for disabling the intel graphics module.

I followed every step in this tutorial (September 2015 version) and installed ElRepo nvidia driver, bumblebee, edited config files etc. as suggested:

https://elrepo.org/tiki/bumblebee

However, now the system freezes at startup. I can only boot the OS at runlevel 3. (The system is a dual boot Win10/Centos)

It seems like bumblebee either fails to start or crashes for some unknown reason. systemctl shows no running bumblebee service, although lsmod shows Nvidia modules running. Calling optirun says bumblebee daemon has either not started or the socket path is incorrect.

Even if I start the bumbleebeed service manually, and then try to go to runlevel 5, the system freezes with no GUI.

Where should I start tracing the issue now?

EDIT Some logs:

systemctl status bumblebeed.service:

Loaded: loaded (/usr/lib/systemd/system/bumblebeed.service; enabled; vendor preset: disabled) Active: Inactive (dead)

manually starting bumblebee by invoking bumblebeed -vv:

[ 1057.431160] [DEBUG]Found card: 01:00.0 (discrete)

[ 1057.431211] [DEBUG]Found card: 00:02.0 (integrated)

[ 1057.431219] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf

[ 1057.431605] [INFO]Configured driver: nvidia

[ 1057.431625] [DEBUG]Skipping auto-detection, using configured driver 'nvidia'

[ 1057.431781] [DEBUG]Process /sbin/modprobe started, PID 2518.

[ 1057.431882] [DEBUG]Hiding stderr for execution of /sbin/modprobe

[ 1057.434369] [DEBUG]SIGCHILD received, but wait failed with No child processes

[ 1057.434484] [DEBUG]bbswitch has been detected.

[ 1057.434493] [INFO]Switching method 'bbswitch' is available and will be used.

[ 1057.434500] [DEBUG]Active configuration:

[ 1057.434505] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf

[ 1057.434511] [DEBUG] X display: :8

[ 1057.434517] [DEBUG] LD_LIBRARY_PATH: /usr/lib64/nvidia:/usr/lib64/vdpau: /usr/lib/nvidia:/usr/lib/vdpau

[ 1057.434523] [DEBUG] Socket path: /var/run/bumblebee.socket

[ 1057.434528] [DEBUG] pidfile: /var/run/bumblebeed.pid

[ 1057.434534] [DEBUG] xorg.conf file: /etc/bumblebee/xorg.conf.nvidia

[ 1057.434539] [DEBUG] xorg.conf.d dir: /etc/bumblebee/xorg.conf.d

[ 1057.434545] [DEBUG] ModulePath: /usr/lib64/xorg/modules/extensions/nvidia,/usr/lib64/xorg/modules

[ 1057.434551] [DEBUG] GID name: bumblebee

[ 1057.434556] [DEBUG] Power method: bbswitch

[ 1057.434562] [DEBUG] Stop X on exit: 1

[ 1057.434568] [DEBUG] Driver: nvidia

[ 1057.434574] [DEBUG] Driver module: nvidia

[ 1057.434579] [DEBUG] Card shutdown state: 1

[ 1057.434725] [DEBUG]Process /sbin/modprobe started, PID 2519.

[ 1057.434817] [DEBUG]Hiding stderr for execution of /sbin/modprobe

[ 1057.437091] [DEBUG]SIGCHILD received, but wait failed with No child processes

[ 1057.437127] [DEBUG]Configuration test passed.

[ 1057.437723] [INFO]bumblebeed 3.2.1 started

[ 1057.438288] [INFO]Unloading nvidia driver

[ 1057.438436] [DEBUG]Process rmmod started, PID 2520.

rmmod: ERROR: Module nvidia is in use by: nvidia_modeset

[ 1057.440466] [DEBUG]Process with PID 2520 returned code 1

[ 1060.455840] [ERROR]Unloading nvidia driver timed out.

[ 1060.455927] [DEBUG]Drivers are still loaded, unable to disable card

[ 1060.455935] [INFO]Initialization completed - now handling client requests

These are running nvidia and i915 kernel modules: lsmod|grep nvidia lsmod|grep i915

nvidia_drm 43405 0

nvidia_modeset 764322 1 nvidia_drm

nvidia 11479193 1 nvidia_modeset

drm_kms_helper 125008 2 i915,nvidia_drm

drm 349210 4 i915,drm_kms_helper,nvidia_drm

i2c_core 40582 8 drm,i915,i2c_i801,i2c_hid,drm_kms_helper,i2c_algo_bit,nvidia,videodev

i915 1097338 0

i2c_algo_bit 13413 1 i915

video 24400 2 i915,asus_wmi

  • lsmod shows only kernel modules and bumblebee doesn't require it's own kernel module (although it needs nvidia and i915 modules to function). To view state of services explore systemctl and journalctl commands - eg. to see if bumblebee is even running: systemctl status bumblebeed.service we'd need to see some relevant log entries to know what's going on. – Marek Rost Sep 10 '16 at 15:12
  • @MarekRost I edited the question and included some logs at runlevel 3. Please let me know if any other diagnostic is needed. – Sourena Sep 10 '16 at 15:55
  • It seems that both daemon and manual startup fail for you. From the log it's apparent that bumblebee is configured to use bbswitch to automatically disable nvidia card until it's needed, but this operation fails to remove the nvidia module from kernel. Looks like your question is in a sense a duplicate of this one: superuser.com/questions/660624/… – Marek Rost Sep 10 '16 at 20:21
  • Modifying /etc/modprobe.d/nvidia.conf did not fix the problem. Upon your guidance I came across this bug report which indicates rmmod cannot efficiently disable nvidia driver module, and the code should be replaced with something like modprobe -r : github.com/Bumblebee-Project/Bumblebee/issues/699 Will I be able to compile this for centos, and do you think it may fix it? github.com/Bumblebee-Project/Bumblebee/pull/762/files – Sourena Sep 10 '16 at 21:22
  • You can try - i suppose another solution could be to switch to nouveau instead of nvidia driver. It does have decent performance on some cards. – Marek Rost Sep 21 '16 at 10:25

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.