[slurm-users] Running vnc after srun fails but works after a direct ssh

[Brian Andrus]

vncserver calls several other things, some of which need a terminal.  
There's more to it, but the bottom line is:  use sbatch rather than srun 
to start a vncserver session.

We merely run vncserver, back-grounding it and then wait.
Something like:

/vncserver :1 &//
//wait//
//echo "user logged out."//
//exit//
/

Simplified, but works well. We can do additional tasks once they start 
the vncserver (eg stage data) and once they log out (clean up files).

Brian Andrus

On 5/15/2021 5:02 AM, Jeremy Fix wrote:
> Hello !
>
> I'm facing a weird issue. With one user, call it gpupro_user , if I log
> with ssh on a compute node, I can run a vncserver (see command [1]
> below) succesfully (in my case, a tigervnc server). However, if I
> allocate the exact same node through a srun (see command [2] below),
> running vnc server fails with the error given at the end of this message.
>
> And finally, if I do the exact same srun, having the exact same computer
> node, from another login (my own login actually), and then start
> vncserver with the exact same command, it works.
>
> So, do you think there is anything in the way we configured the user
> gpupro_user, or maybe declared it in sacctmgr or somewhere that could
> explain why running vncserver from within the srun session fails ?
>
> Thank you for your help.
>
> Have a nice day ;
>
> Jeremy.
>
> [1] vncserver -SecurityTypes None -depth 32 -geometry 1680x1050
>
> [2]  srun --nodelist=tx01 -N 1 -p gpue60 -t 0:30:00 --pty bash
>
> ------- VNC error ----------------------------------------
>
> Please be aware that you are exposing your VNC server to all users on the
> local machine. These users can access your server without authentication!
>
> New 'tx01:1 (gpuaut_2)' desktop at :1 on machine tx01
>
> Starting applications specified in /etc/X11/Xvnc-session
> Log file is /usr/users/gpuaut/gpuaut_2/.vnc/tx01:1.log
>
> Use xtigervncviewer -SecurityTypes None :1 to connect to the VNC server.
>
>
> vncserver: Failed command '/etc/X11/Xvnc-session': 256!
>
> =================== tail -15 /usr/users/gpuaut/gpuaut_2/.vnc/tx01:1.log
> ===================
> Killing Xtigervnc process ID 31975... which seems to be deadlocked.
> Using SIGKILL!
>
> Xvnc TigerVNC 1.7.0 - built Dec  5 2017 09:25:01
> Copyright (C) 1999-2016 TigerVNC Team and many others (see README.txt)
> See http://www.tigervnc.org for information on TigerVNC.
> Underlying X server release 11905000, The X.Org Foundation
>
>
> Sat May 15 13:57:35 2021
>   vncext:      VNC extension running!
>   vncext:      Listening for VNC connections on local interface(s), port 5901
>   vncext:      created VNC server for screen 0
> XIO:  fatal IO error 11 (Resource temporarily unavailable) on X server ":1"
>        after 175 requests (175 known processed) with 0 events remaining.
> Killing Xtigervnc process ID 7169... which seems to be deadlocked. Using
> SIGKILL!
>
> ===========================================================================================
>
> Starting applications specified in /etc/X11/Xvnc-session has failed.
> Maybe try something simple first, e.g.,
>      tigervncserver -xstartup /usr/bin/xterm
> ------- VNC error ----------------------------------------
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20210515/6b60e6dd/attachment.htm>