vino-server eats a lot of cpu time with d-bus requests

Bug #371356 reported by Paolo Benvenuto
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
vino (Ubuntu)
Incomplete
Low
Unassigned

Bug Description

Binary package hint: xserver-xorg-video-intel

Since I upgraded to jaunty, the cpu is busy at 50% on my core2 duo system, mainly due to xserver and dbus-deamon: multiload applets shows a systematic 50% processor use, even if no user app is active.

Even if I remove multiload applet and close gnome-system-monitor, top keep showing a lot of use by xorg and dbus-daemon.

$ cat /proc/version
Linux version 2.6.28-11-server (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #42-Ubuntu SMP Fri Apr 17 02:48:10 UTC 2009

$ lspci
...
00:02.0 VGA compatible controller: Intel Corporation 82G33/G31 Express Integrated Graphics Controller (rev 10)
...

$ sudo lshw
...
     *-pci
          description: Host bridge
          product: 82G33/G31/P35/P31 Express DRAM Controller
          vendor: Intel Corporation
          physical id: 100
          bus info: pci@0000:00:00.0
          version: 10
          width: 32 bits
          clock: 33MHz
          configuration: driver=agpgart-intel module=intel_agp
        *-display UNCLAIMED
             description: VGA compatible controller
             product: 82G33/G31 Express Integrated Graphics Controller
             vendor: Intel Corporation
             physical id: 2
             bus info: pci@0000:00:02.0
             version: 10
             width: 32 bits
             clock: 33MHz
             capabilities: msi pm bus_master cap_list
             configuration: latency=0
...

some top screens:

Tasks: 176 total, 2 running, 174 sleeping, 0 stopped, 0 zombie
Cpu(s): 37.7%us, 18.1%sy, 0.6%ni, 43.3%id, 0.1%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 3346308k total, 3263076k used, 83232k free, 291328k buffers
Swap: 7815612k total, 5272k used, 7810340k free, 1713364k cached

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 3800 root 20 0 460m 148m 11m S 16 4.6 305:42.39 Xorg
 4283 paolo 20 0 3708 1888 652 S 12 0.1 304:07.33 dbus-daemon
 4500 paolo 20 0 195m 182m 8908 S 6 5.6 120:51.67 indicator-apple
11794 paolo 20 0 22896 7280 5820 R 4 0.2 0:00.02 vino-server
 4133 paolo 20 0 26964 7504 5816 S 2 0.2 156:11.59 gnome-session
 4349 paolo 20 0 38616 22m 12m S 2 0.7 55:57.82 python
 4383 paolo 20 0 25888 9200 6192 S 2 0.3 20:52.18 gnome-power-man
 4515 paolo 20 0 27912 14m 10m S 2 0.4 22:05.07 fast-user-switc

Tasks: 176 total, 4 running, 172 sleeping, 0 stopped, 0 zombie
Cpu(s): 37.7%us, 18.1%sy, 0.6%ni, 43.3%id, 0.1%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 3346308k total, 3266304k used, 80004k free, 291456k buffers
Swap: 7815612k total, 5272k used, 7810340k free, 1715548k cached

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 3800 root 20 0 460m 148m 11m S 14 4.6 305:51.98 Xorg
 4283 paolo 20 0 3708 1888 652 S 12 0.1 304:17.06 dbus-daemon
 4133 paolo 20 0 26964 7504 5816 R 6 0.2 156:15.59 gnome-session
 6942 paolo 20 0 424m 162m 31m S 6 5.0 123:47.75 epiphany-browse
 4500 paolo 20 0 195m 182m 8908 S 4 5.6 120:56.32 indicator-apple
 4328 paolo 20 0 53828 25m 14m S 2 0.8 49:48.27 gnome-panel
 4369 paolo 20 0 28324 10m 4680 S 2 0.3 19:39.23 trackerd
 4374 paolo 20 0 29092 15m 11m S 2 0.5 15:26.44 update-notifier
 4383 paolo 20 0 25888 9200 6192 S 2 0.3 20:52.72 gnome-power-man
 4512 paolo 20 0 27348 13m 9404 S 2 0.4 10:30.86 gnome-keyboard-
 4515 paolo 20 0 27912 14m 10m S 2 0.4 22:05.58 fast-user-switc
22231 paolo 20 0 450m 145m 19m S 2 4.4 40:16.65 rhythmbox

Tasks: 176 total, 3 running, 173 sleeping, 0 stopped, 0 zombie
Cpu(s): 34.8%us, 18.0%sy, 0.3%ni, 46.6%id, 0.2%wa, 0.0%hi, 0.2%si, 0.0%st
Mem: 3346308k total, 3266128k used, 80180k free, 291480k buffers
Swap: 7815612k total, 5272k used, 7810340k free, 1715016k cached

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 3800 root 20 0 460m 148m 11m S 8 4.6 305:54.70 Xorg
 4283 paolo 20 0 3708 1888 652 S 7 0.1 304:19.11 dbus-daemon
 4500 paolo 20 0 195m 182m 8908 S 6 5.6 120:57.29 indicator-apple
 4133 paolo 20 0 26964 7504 5816 S 4 0.2 156:16.45 gnome-session
 6942 paolo 20 0 424m 162m 31m S 3 5.0 123:48.71 epiphany-browse
 4383 paolo 20 0 25888 9200 6192 S 2 0.3 20:52.88 gnome-power-man
22231 paolo 20 0 450m 145m 19m S 2 4.4 40:17.06 rhythmbox
 4291 paolo 20 0 9368 5828 2296 S 1 0.2 136:11.99 gconfd-2
 4288 paolo 20 0 228m 5588 4144 S 1 0.2 36:59.11 pulseaudio
 4309 paolo 20 0 40280 10m 7436 S 1 0.3 49:22.35 gnome-settings-
 4349 paolo 20 0 38616 22m 12m S 1 0.7 55:59.51 python
 4437 paolo 39 19 28692 10m 4648 S 1 0.3 19:17.22 tracker-indexer
    7 root 15 -5 0 0 0 S 1 0.0 4:25.28 ksoftirqd/1
 4316 paolo 20 0 5952 2284 1804 S 1 0.1 10:31.91 gvfsd
 4366 paolo 20 0 69036 11m 9268 S 1 0.4 13:03.78 evolution-alarm
 4369 paolo 20 0 28324 10m 4680 S 1 0.3 19:39.37 trackerd
32064 paolo 20 0 34360 14m 9680 R 1 0.5 0:03.59 gnome-terminal

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :
Revision history for this message
Alan Herrlich (alan-herrlich) wrote :

I have a similar issue which I have tracked down to vino-server or at least disabling remote desktop sharing eliminates the problem. I see that you are also running vino-server. I hunted it down using dbus-monitor, this command will show who is sending dbus messages...

dbus-monitor | grep 'string \"[a-zA-Z\.]*\.[a-zA-Z\.]*\"'

I found "org.freedesktop.indicator" and "org.gnome.Vino" were sending a lot of messages. When I unclicked "Sharing" -> "Allow other users to view your desktop" the dbus messages stopped and cpu utilization returned to normal. Turning vino back on causes the problem again.

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

I tried what you are suggesting, but in my case even if I disable "Allow other users to view your desktop" the cpu keeps being intensively used.

I confirm that

dbus-monitor | grep 'string \"[a-zA-Z\.]*\.[a-zA-Z\.]*\"'

gives same output as you're saying, even if I disable "Allow other users to view your desktop"

Revision history for this message
Alan Herrlich (alan-herrlich) wrote :

Strange... maybe your system is setup differently than mine. Within a few seconds of me changing the setting the dbus-monitor command quiets down and CPU usage goes down. You're clicking on the top level "...view your desktop" not the lower "...control your desktop" setting? Mine show the message "Nobody can access your desktop". You could also try

killall vino-server

or try uninstalling the vino package in the Synaptic Package Manager. I suppose it might need a reboot or something to kill vino and stop those dbus messages.

Revision history for this message
Bryce Harrington (bryce) wrote :

Hi donpaolo,

Thanks for including an image to demonstrate the issue. Could you also please attach the output of `lspci -vvnn`, and attach your /var/log/Xorg.0.log or Xorg.0.log.old file from after reproducing this issue. If you've made any customizations to your /etc/X11/xorg.conf please attach that as well.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

tags: added: needs-xorglog
Changed in xserver-xorg-video-intel (Ubuntu):
status: New → Incomplete
Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

I realized eventually that I (supposedly) had various vino-server running. I'm saying supposedly beacuse I didn't see them with ps. After various "killall vino-server" the cpu usage ended, and now everything is normal.

I'm going to see if the defect reappears when I connect to vino-server again.

Bryce Harrington (bryce)
tags: added: high-cpu
Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

supposedly its a problem with vino-server, not with the intel driver.

I see the 50% cpu usage when I reboot, and killing repeteadly vino-server solves the issue.

Changing bug title and package accordingly.

affects: xserver-xorg-video-intel (Ubuntu) → vino (Ubuntu)
Changed in vino (Ubuntu):
status: Incomplete → New
summary: - Intel VGA controller 82G33/G31 keeps processor always busy
+ vino-server eats a lot of cpu time with d-bus requests
Revision history for this message
Sebastien Bacher (seb128) wrote :

could be that one instance is running and that gnome-session tries to start in loop an another one, do you have it registered in your session?

Changed in vino (Ubuntu):
importance: Undecided → Low
status: New → Incomplete
Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

How do I check it?

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

I have vino-server in my gnome startup applications, but I didn't put it there myself

Revision history for this message
Alan Herrlich (alan-herrlich) wrote :

FYI: I removed "Remote Desktop Sharing" from System -> Preferences -> Startup Applications. I did killall vino-server a few times but it stayed around. And then I was able to "Allow other users to view your desktop" and my CPU usage issue stayed low and normal. Removing it from the startup seems to have fixed it. Like Paolo I did not myself put Remote Desktop in my startup. My machine was a machine freshly installed with Ubuntu 8.? and recently upgraded to 9.04.

Revision history for this message
Alessandro Morgantini (gpz500) wrote :

For me it is the same as Alan Herrlich.
But, in a freshly created new user, "Remote Desktop Sharing" in Startup Applications is enabled (by default) and everything seems to be OK.
So, I think the problem is related to the upgrade process.

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

Since last comment I purged and installed again vino

Now at restarting the system I have 2 instances of vino-server:

$ ps ax|grep vino
 3599 ? S 0:00 /usr/lib/vino/vino-server
 5908 ? D 0:00 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-v3yT2x/ --sm-client-id 1085af4c166cc32bb1124262707933240600000041570007 --screen 0
 5910 pts/0 R+ 0:00 grep vino

and now killall -9 vino-server stops vino from eating cpu time.

Unfortunatly after a reboot I have again the 2 instances of vino-server and the consequent cpu-eating situation

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

If I disable "remote desktop" from gnome startup application, vino-server isn't running any more.

So it seem that the two instances of vino-server are launched from there. Isn't it weird?

Revision history for this message
Alessandro Morgantini (gpz500) wrote :

Ok, it seems that the starting method of vino-server is changed with Jaunty, or something similar.
With the new method, the vino-server process (without any arguments) is always started by Startup Applications, even if the option in System->Preferences->Remote Desktop is disabled; enabling this just instructs the existing vino-server to listen to the 5900 port and accept connections.
With the old method, I think, System->Preferences->Remote Desktop used to enable its own vino-server process (the process with arguments).
Because of a bug in the upgrade procedure, for the old users, both of the processes are there, causing the problem with the cpu.

Comparing a new user created after the upgrade to Jaunty (not presenting any bug) and the old one, created before the upgrade (with the bug), I found a feasable workaround. To adopt the new method:
1) delete the .desktop file which refers to vino-server in the folder ~/.config/gnome-session/saved-session; and
2) enable "Remote Desktop" in Startup Applications;
3) restart the session (logout & login).

Revision history for this message
Paolo Benvenuto (donpaolo) wrote :

I followed the instructions in https://bugs.launchpad.net/ubuntu/+source/vino/+bug/371356/comments/15 , but unfortunatly it doesn't work. At reboot I still have another vino-server instance trying to run, as you can see in this ps's executed few seconds each after the other:

$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8313 ? R 0:00 [vino-server]
 8315 pts/0 S+ 0:00 grep vino
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8535 ? R 0:00 /usr/lib/vino/vino-server
 8537 pts/0 R+ 0:00 grep vino
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8594 ? Z 0:00 [vino-server] <defunct>
 8596 pts/0 R+ 0:00 grep vino
 8597 ? R 0:00 /usr/lib/vino/vino-server
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8750 ? R 0:00 /usr/lib/vino/vino-server
 8752 pts/0 R+ 0:00 grep vino
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8785 ? R 0:00 /usr/lib/vino/vino-server
 8787 pts/0 R+ 0:00 grep vino
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8818 ? R 0:00 /usr/lib/vino/vino-server
 8820 pts/0 R+ 0:00 grep vino
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8855 pts/0 R+ 0:00 grep vino
 8856 ? R 0:00 /usr/lib/vino/vino-server
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8888 pts/0 R+ 0:00 grep vino
 8889 ? R 0:00 /usr/lib/vino/vino-server
paolo@paolo:~$ ps ax|grep vino
 3607 ? S 0:01 /usr/lib/vino/vino-server --sm-config-prefix /vino-server-lJFb2o/ --sm-client-id 10c5e3eb189b6c3322124263020744463500000058740007 --screen 0
 8925 ? Z 0:00 [vino-server] <defunct>
 8927 ? R 0:00 /usr/lib/vino/vino-server
 8928 pts/0 R+ 0:00 grep vino

Process 3067 is there, the other vino-server runs and dies, runs and dies, etc.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.