Execute V-REP without GUI via ssh

Report crashes, strange behaviour, or apparent bugs
coppelia
Site Admin
Posts: 6224
Joined: 14 Dec 2012, 00:25

Re: Execute V-REP without GUI via ssh

Post by coppelia » 15 Jul 2016, 08:53

Jessica,

before transmitting it, you could save it to file (simSaveImage) to make sure that the vision sensor is operational. Remember that you should do this in the sensing section of a child script, otherwise you will retrieve the image from previous simulation step (and no image at all in the first simulation step). Unless you explicitely handle your vision sensor.

There is an example how to publish/subscribe to an image with the new RosInterface: rosInterfaceTopicPublisherAndSubscriber.ttt

Cheers

nehagarg
Posts: 17
Joined: 06 Mar 2015, 10:54

Re: Execute V-REP without GUI via ssh

Post by nehagarg » 22 Dec 2016, 12:01

Hi

I am also trying to run vrep in headless mode without GUI. I followed the solution in the post with xvfb-run and when I run the following command

xvfb-run --auto-servernum --server-num=1 -s "-screen 0 640x480x24" ./vrep.sh -h <scene name>

vrep is able to load all the plugins and start.

However as soon as I start simulation through ros api , vrep crashes with error: "Could not initialize OpenGL Aborted(Core Dumped)"

If I issue command xvfb-run --auto-servernum --server-num=1 -s "-screen 0 640x480x24" ./vrep.sh -h -s5000 -q <scene name>, then also vrep crashes immediately with the same error.

Is there a way to fix it?

Thanks.

Neha

onlytailei
Posts: 4
Joined: 23 Mar 2017, 17:02

Re: Execute V-REP without GUI via ssh

Post by onlytailei » 24 Mar 2017, 10:43

Hi,

Would you mind to check this problem?
viewtopic.php?f=9&t=5606

Code: Select all

xvfb-run --auto-servernum --server-num=1 -s "-screen 0 640x480x24" ./vrep.sh -h -s -q my-sim.ttt
Do not work for vision sensor.

coppelia
Site Admin
Posts: 6224
Joined: 14 Dec 2012, 00:25

Re: Execute V-REP without GUI via ssh

Post by coppelia » 25 Mar 2017, 12:06

When you start V-REP in headless mode, you basically run the exact same binaries as the V-REP with GUI. It is not a true headless version of V-REP. I see currently two possibilities:
  • You can try to run with different graphic settings and see if this makes a difference. Have a look at file system/usrset.txt. In there, try out different values for offscreenContextType, fboType, forceFboViaExt and vboOperation.
  • You can recompile V-REP in true headless mode by compiling V-REP with the makefile_noGui_noGl makefile. In that case however, vision sensors will only work if they rely on a plugin for image generation (e.g. v_repExtPovRay).
Cheers

nng
Posts: 2
Joined: 31 Oct 2014, 11:51
Contact:

Re: Execute V-REP without GUI via ssh

Post by nng » 12 Oct 2017, 14:19

The solution of rezama unfortunately only seems to work on remote computers with an X server.
Even having libgl1-mesa-dri installed does not work
xvfb-run --auto-servernum --server-num=1 -s "-screen 0 640x480x24" ./vrep.sh -h

is there any plan to provide a real headless version of VREP?
If yes what is the time line for this?
This issue was originally posted in Feb 2014!!!

coppelia
Site Admin
Posts: 6224
Joined: 14 Dec 2012, 00:25

Re: Execute V-REP without GUI via ssh

Post by coppelia » 13 Oct 2017, 07:30

This is currently not high priority. But the plan is to replace the whole graphics engine with something that is more detachable, scalable and maintainable. But we can't make any fixed schedule for that..

Cheers

Post Reply