Hello! I have been trying to get Chimera working through ssh, where the Chimera installation is located on a Red Hat based cluster, but the user is using a MacOS machine to access the cluster. Looking through the archives of the mailing list, I found a suggestion to use VirtualGL to run Chimera through ssh. VirtualGL seems to be running as intended, but isn't working entirely correctly with Chimera (it runs glxgears and other applications fine). It gets farther in start up than without VirtualGL, but crashes on the Initializing OpenGL step around 80% of the time. I have traced the point of failure to reside within a call to the makeGraphicsWindow function within the tkgui.py file. Specifically, it fails at the line within the makeGraphicsWindow function: graphics = Togl.Togl(parent, **kw) which is found somewhere around line 2690 within the try / except portion which is itself nested in the while 1 loop. I can't determine why exactly it is failing here, and trying to trace the problem to a specific line within the instantiation of the Togl object proved fruitless. I was wondering if anyone would be able to shed some light on what could be causing this issue, or know how to fix it? Alternatively, if anyone knows of another good solution for using Chimera through ssh on a Mac machine, that information would be greatly appreciated. Thanks for taking the time to read my inquiry! Best, Logan Sweezy Scientific Systems Admin Rockefeller University
To debug bug this, you'll need to use gdb and see which glX library call is giving you problems. Make a copy of the CHIMERA/bin/chimera shell script in the same directory and have it invoke 'gdb "$python"' and then within gdb start chimera with two commands: break _exit run $CHIMERA/share/__main__.py --debug The --debug should put the X11 connection into synchronous mode so when the error occurs, it should be attributed to the correct glX library call. Then give the gdb command bt to show the backtrace. It would also help if VirtualGL had a way to log the arguments to the glX calls it intercepts. I see that VirtualGL is being actively developed, so make sure your VirtualGL is up to date. HTH, Greg On 06/28/2018 02:44 PM, Logan Sweezy wrote:
Hello!
I have been trying to get Chimera working through ssh, where the Chimera installation is located on a Red Hat based cluster, but the user is using a MacOS machine to access the cluster. Looking through the archives of the mailing list, I found a suggestion to use VirtualGL to run Chimera through ssh. VirtualGL seems to be running as intended, but isn't working entirely correctly with Chimera (it runs glxgears and other applications fine). It gets farther in start up than without VirtualGL, but crashes on the Initializing OpenGL step around 80% of the time. I have traced the point of failure to reside within a call to the makeGraphicsWindow function within the tkgui.py file. Specifically, it fails at the line within the makeGraphicsWindow function:
graphics = Togl.Togl(parent, **kw)
which is found somewhere around line 2690 within the try / except portion which is itself nested in the while 1 loop. I can't determine why exactly it is failing here, and trying to trace the problem to a specific line within the instantiation of the Togl object proved fruitless. I was wondering if anyone would be able to shed some light on what could be causing this issue, or know how to fix it?
Alternatively, if anyone knows of another good solution for using Chimera through ssh on a Mac machine, that information would be greatly appreciated.
Thanks for taking the time to read my inquiry!
Best, Logan Sweezy
/Scientific Systems Admin/
/Rockefeller University /
//
_______________________________________________ Chimera-users mailing list: Chimera-users@cgl.ucsf.edu Manage subscription: http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
participants (2)
-
Greg Couch -
Logan Sweezy