Please read this file carefully, if you would like to take advantage of mental
ray DBR and V-Ray DR with Qube!
All worker machines need to have the following installed and working
- Qube core 5.1 or above
- Qube Worker 5.1 or above
- Qube 3dsmax Jobtype 5.2 or above
- 3dsmax 8.0 or above
First and foremost, mental ray DBR or V-Ray DR for 3dsmax must be setup, and
tested to be working properly, outside of Qube! Please consult your 3dsmax
documentation for details on setting it up.
Because of the way mental ray DBR is implemented on 3dsmax, the jobtype
needs to be able to write to the "max.rayhosts" file found in the
"mentalray" subfolder, in the system's 3dsmax installation folder (by
default, "C:/Program Files/Autodesk/3ds max 9/mentalray"). Accordingly,
you'll need to make sure that all execution users on all workers have
write-access to the file and containing folder.
Similarly for V-Ray DR, the jobtype needs to be able to write to the
"vray_dr.cfg" file found in the system's 3dsmax installation "plugcfg"
folder, which is something like:
"C:\Users\USERNAME\AppData\Local\Autodesk\3dsMax\2011 - 64bin\enu]\plugcfg"
Accordingly, you'll need to make sure that all execution users on all
workers have write-access to the file and containing folder.
Open or create a scene.
Make sure your scene is properly set up to use mental ray or V-Ray as the
"Rendering" menu -> "Render" -> "Common" tab -> "Assign Renderer" rollout
Save the scene if changes were made.
Open the Qube! Submission dialog: "Qube!" Menu -> "Submit Render..."
In the "Parameters" section, select "mentalray" or "vray" in the "Enable
In the "Qube Job Basics" section, specify the desired number of Subjob
Set up other parameters as necessary.
Submit the render, by clicking the "Submit" button.
The jobtype will re-write the [max.rayhosts] file in the following
location on the master node (i.e., where subjob 0 runs):
The jobtype will try to back up the original rayhosts file, if it
exists, before the execution of the job, and then restore it
afterwards, but if the job dies in the middle, the restore may not
[Known Issues and Limitations]
Workers will wait idle until the specified number of CPUs are
collected for the job. For example, if a job specified 8 CPUs, the
job will not start processing until all 8 CPUs are available--
sometimes the first 7 CPUs can wait for a long time before the 8th
one becomes available.
If renderers other than mental ray or V-Ray are designated, the job will
render the only using the master node-- other reserved nodes will sit idle
until the job is finished or killed.
Due to the inherent dependency on "max.rayhosts" file of mental ray
satellite renders, attempts to run more than one satellite job
concurrently on the same machine as a master node may cause serious
instability in the system, and thus is unsupported.
All limitations of the 3dsmax jobtype apply to satellite rendering
on Qube. For example, the scenefile must be located on a shared
file-server path where all workers may uniformly access.
All limitations of mental ray and mental ray DBR for 3dsmax, V-Ray and V-Ray
DR for 3dsmax, inherently apply. For example, if a slave node becomes
inaccessible during a render, the entire job may crash, due to the nature of
the rendering software.