then i logged in using vglconnect -x user:pass as described. Then in that same ssh session , i did vglrun glxgears, and it just wont run . This is the output of vglrun +v glxgears :
[VGL] Shared memory segment ID for vglconfig: 360448 [VGL] VirtualGL v2.3.2 64-bit (Build 20121002) [VGL] Opening local display :0
Error: couldn't open display 129.21.147.40:0
Somebody help :/
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Did you follow the "Sanity Check" procedure described in the User's Guide (https://cdn.rawgit.com/VirtualGL/virtualgl/2.5.1/doc/index.html#hd006, under "Sanity Check")?
Last edit: DRC 2016-10-31
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am sorry for the late reply.
I boot into the console , and when I want , i do sudo kdm to start a display manager.
I ran sudo /opt/VirtualGL/bin/vglserver_config
the output is :
Creating vglusers group ...
groupadd: group 'vglusers' already exists
Could not add vglusers group (probably because it already exists.)
... Granting read permission to /etc/opt/VirtualGL/ for vglusers group ...
... Creating /etc/modprobe.d/virtualgl.conf to set requested permissions for
/dev/nvidia* ...
... Attempting to remove nvidia module from memory so device permissions
will be reloaded ...
rmmod: ERROR: Module nvidia is in use by: nvidia_modeset
... Modifying /etc/X11/xorg.conf to enable DRI permissions for
vglusers group ...
... Adding vglgenkey to /etc/kde4/kdm/Xsetup script ...
... Adding display-setup-script=vglgenkey to /etc/lightdm/lightdm.conf ...
... Enabling XTEST extension in /etc/kde4/kdm/kdmrc ...
Done. You must restart the display manager for the changes to take effect.
IMPORTANT NOTE: Your system uses modprobe.d to set device permissions. You
must execute rmmod nvidia with the display manager stopped in order for the
new device permission settings to become effective.
Notice the error.
How do I proceed ?
Thanks a ton for your help!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
That error is normal and is a result of you ignoring the instructions in the User's Guide that said to stop the display manager before running vglserver_config. The output of the script tells you exactly what to do to overcome it: stop the display manager and execute rmmod nvidia. I would suggest that you do the following:
Stop the display manager
Run rmmod nvidia
Re-run vglserver_config
Restart the display manager
If VirtualGL still doesn't work, then I would suggest editing /etc/kde4/kdm/Xsetup and changing the vglgenkey line such that the output of vglgenkey is redirected to a file:
{path_to}/vglgenkey >/tmp/out 2>&1
Then stop/restart the display manager and see if /tmp/out contains any obvious errors.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
That could be why VirtualGL isn't working. It assumes that you are running a display manager. The purpose of vglserver_config is to set up the display manager such that it will grant access to VirtualGL while the system is sitting at the login prompt.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I was able to get glxgears to display after installing xdm. But after a reboot when I do it , I am getting:
Invalid MIT-MAGIC-COOKIE-1 key[VGL] ERROR: Could not open display :0
If i reinstall again , then do vglconnect , it works . And then after reboot , i get the error i mentioned.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Let's focus on KDM first. Please follow the diagnostic procedure I outlined above so we can figure out whether vglgenkey is completing successfully. If it isn't, then there is no way that you'll ever be able to successfully use VirtualGL.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
After installing xdm , i am able to run virtual gl . Thanks!
I login using vglconnect , and when I do vglrun glxgears , it shows up.
All I want to do now is be able to run a long going task , and be able to 'check in' on it to see the progress.
So for eg, if i run vglrun glxgears, i want to 'detach' (much like screen) , and have it continue running in the server. And then at an arbitrary amount of time later, i want to reattach back.
What should I do ?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Use TurboVNC. Start a TurboVNC session, connect to it with the TurboVNC Viewer, start your 3D application using vglrun inside of the TurboVNC session, then you can disconnect and leave the TurboVNC session running for as long as you want.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks !!.
Whatever I wanted is more or less accomplished.
I just dont want to waste my resources starting up a desktop.
is it possible to use turbovnc to just see one (or several) windows without the whole desktop?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
"Seamless windows" (displaying one client-side window for every application window) is unfortunately not possible with any VNC solution, due to limitations in the protocol. However, you can greatly reduce the resource usage by using a stripped-down window manager such as OpenBox or TWM. You can also configure OpenBox to isolate individual applications so that only those applications are accessible within the TurboVNC session:
If you need seamless windows, you might also look at Xpra. I don't have a lot of experience with it, and I doubt the performance will be as good as TurboVNC, but if my understanding is correct, it should also be a stateless solution (meaning you can disconnect from and reconnect to a running application.)
Last edit: DRC 2016-12-05
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hey , i just noticed something.
When I am running my long task via turbovnc , and I login from kdm(thats my display manager) via gnome , and after I log out , the job that was running earlier just crashes.
Is this documented behaviour , or am I doing something wrong ?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If you log into Display :0 and log out again, it resets the display and disconnects any connected X clients. Since VGL is forwarding GLX calls to Display :0, it is an X client of that display. If you need to use the root display while using VGL, just leave the root display logged in. Or find me a financial sponsor to implement this:
I am trying to get opengl work for one of my projects.
I have done
/opt/VirtualGL/bin/vglserver_config
then chose all recommended settings
then i logged in using vglconnect -x user:pass as described. Then in that same ssh session , i did vglrun glxgears, and it just wont run . This is the output of vglrun +v glxgears :
[VGL] Shared memory segment ID for vglconfig: 360448
[VGL] VirtualGL v2.3.2 64-bit (Build 20121002)
[VGL] Opening local display :0
Error: couldn't open display 129.21.147.40:0
Somebody help :/
Last edit: DRC 2016-10-31
I am sorry for the late reply.
I boot into the console , and when I want , i do sudo kdm to start a display manager.
I ran sudo /opt/VirtualGL/bin/vglserver_config
the output is :
Creating vglusers group ...
groupadd: group 'vglusers' already exists
Could not add vglusers group (probably because it already exists.)
... Granting read permission to /etc/opt/VirtualGL/ for vglusers group ...
... Creating /etc/modprobe.d/virtualgl.conf to set requested permissions for
/dev/nvidia* ...
... Attempting to remove nvidia module from memory so device permissions
will be reloaded ...
rmmod: ERROR: Module nvidia is in use by: nvidia_modeset
... Modifying /etc/X11/xorg.conf to enable DRI permissions for
vglusers group ...
... Adding vglgenkey to /etc/kde4/kdm/Xsetup script ...
... Adding display-setup-script=vglgenkey to /etc/lightdm/lightdm.conf ...
... Enabling XTEST extension in /etc/kde4/kdm/kdmrc ...
Done. You must restart the display manager for the changes to take effect.
IMPORTANT NOTE: Your system uses modprobe.d to set device permissions. You
must execute rmmod nvidia with the display manager stopped in order for the
new device permission settings to become effective.
Notice the error.
How do I proceed ?
Thanks a ton for your help!
That error is normal and is a result of you ignoring the instructions in the User's Guide that said to stop the display manager before running vglserver_config. The output of the script tells you exactly what to do to overcome it: stop the display manager and execute
rmmod nvidia
. I would suggest that you do the following:rmmod nvidia
If VirtualGL still doesn't work, then I would suggest editing /etc/kde4/kdm/Xsetup and changing the vglgenkey line such that the output of vglgenkey is redirected to a file:
{path_to}/vglgenkey >/tmp/out 2>&1
Then stop/restart the display manager and see if /tmp/out contains any obvious errors.
what is my display manager ?
I am not automatically running kdm, gdm or anything
That could be why VirtualGL isn't working. It assumes that you are running a display manager. The purpose of vglserver_config is to set up the display manager such that it will grant access to VirtualGL while the system is sitting at the login prompt.
I was able to get glxgears to display after installing xdm. But after a reboot when I do it , I am getting:
Invalid MIT-MAGIC-COOKIE-1 key[VGL] ERROR: Could not open display :0
If i reinstall again , then do vglconnect , it works . And then after reboot , i get the error i mentioned.
Let's focus on KDM first. Please follow the diagnostic procedure I outlined above so we can figure out whether vglgenkey is completing successfully. If it isn't, then there is no way that you'll ever be able to successfully use VirtualGL.
After installing xdm , i am able to run virtual gl . Thanks!
I login using vglconnect , and when I do vglrun glxgears , it shows up.
All I want to do now is be able to run a long going task , and be able to 'check in' on it to see the progress.
So for eg, if i run vglrun glxgears, i want to 'detach' (much like screen) , and have it continue running in the server. And then at an arbitrary amount of time later, i want to reattach back.
What should I do ?
Use TurboVNC. Start a TurboVNC session, connect to it with the TurboVNC Viewer, start your 3D application using
vglrun
inside of the TurboVNC session, then you can disconnect and leave the TurboVNC session running for as long as you want.Thanks !!.
Whatever I wanted is more or less accomplished.
I just dont want to waste my resources starting up a desktop.
is it possible to use turbovnc to just see one (or several) windows without the whole desktop?
"Seamless windows" (displaying one client-side window for every application window) is unfortunately not possible with any VNC solution, due to limitations in the protocol. However, you can greatly reduce the resource usage by using a stripped-down window manager such as OpenBox or TWM. You can also configure OpenBox to isolate individual applications so that only those applications are accessible within the TurboVNC session:
http://www.turbovnc.org/Documentation/ApplicationIsolation
If you need seamless windows, you might also look at Xpra. I don't have a lot of experience with it, and I doubt the performance will be as good as TurboVNC, but if my understanding is correct, it should also be a stateless solution (meaning you can disconnect from and reconnect to a running application.)
Last edit: DRC 2016-12-05
thanks a lot!
you have been very kind and helpful
have a nice day!
I cant find how to change vncserver to load up twm instead of kdm :/
Just edit ~/.vnc/xstartup.turbovnc and delete everything except these lines near the bottom:
Last edit: DRC 2016-12-05
Another option is to leave xstartup.turbovnc alone, copy those same lines into ~/.Xclients, then:
thanks a lot for that superfast reply.
though adding the file Xclients didnt work for me.
The first method (editing xstartup.turbovnc) worked
There's a dot in the filename. It's ".Xclients", not "Xclients".
Hey , i just noticed something.
When I am running my long task via turbovnc , and I login from kdm(thats my display manager) via gnome , and after I log out , the job that was running earlier just crashes.
Is this documented behaviour , or am I doing something wrong ?
If you log into Display :0 and log out again, it resets the display and disconnects any connected X clients. Since VGL is forwarding GLX calls to Display :0, it is an X client of that display. If you need to use the root display while using VGL, just leave the root display logged in. Or find me a financial sponsor to implement this:
https://github.com/VirtualGL/virtualgl/issues/10