This is an Open OnDemand application for rendering animation frames. We currently support AutoDesk's Maya as its rendering engine.
First you need to upload your project to the cluster if the files are not already there. An easy way may be the File application provided by OnDemand. In it you can simply drag and drop a directory if it's less than 10 GB.
You'll want to upload it to your $HOME/maya/projects directory. In this example $HOME is
/home/jeff and it will change depending on the site and your username. You may have to
make this directory, and how to do that is also highlighted in this picture in the upper right.
If you're using WinSCP, here are is a basic tutorial on how to connect to your site.
Create a new connection like I have here. scp.osc.edu is OSC's ftp server, your site's hostname
will differ if you don't use OSC. You'll have to specify your own username & password in the
appropriate fields fields.
Open this connection up and navigate to $HOME/maya/projects. In this example $HOME is
is /users/PZS0714/johrstrom. Again, you may have to make the sub-directories, but the buttons
are shown.
Once you're connected, you can simply drag and drop folders back and forth.
Click the 
button to create a new project.
There you'll need to give it a name and optionally a description. You'll also need to specify
where this project directory is. It should be the location of the files you just uploaded.
Use the file picker 
to help choose
this directory. Press 
and you're ready to
get started.
Job settings are essentially a set of arguments you want to execute a job with. Logically, it's something like I want to render the first 10 frames of this scene and get an email when it's done.
To create set of job settings simply press the 
button. Fill out all the nessecary form items. The form items are detailed below.
| Option | Description | Example |
|---|---|---|
| name | The name of the settings. | always render first 10 |
| frames | The frames you want to render. | 1-10 |
| camera | The camera view you want to render. | camera1 |
| file | The ma or mb file you want to render. | /home/me/maya/projects/project_1/scenes/testScene1.ma |
| chargeback project | The project you're charging too. You only have to specify this if you're in multiple projects. | PZS0714 |
| cluster | The cluster you want to submit to. | owens |
| renderer | The rendering engine you want to use. | arnold |
| additional command line options | Extra options you want to give to the command. See more here. | -verb -b 1 -ai:lve 0 |
| walltime | The amount of time in hours your job will be scheduled. | 3 |
| email when finished | A choice to email you when the job is complete. | (checked) |
| skip existing | A choice for the renderer to skip rendered images if they already exist. | (not checked) |
Press 
and you're all done.
Once you've created the settings you can always go back and edit it through the 
button.
And, for whatever reason, you decide you don't want it anymore simply press the 
button
to delete it.
Now that you have a job all configured you're ready to submit it. to do so, Simply press the
button.
You should see a submission similar to this one below
turn it's status 
, get a job id, and finally running like this job that's
running:
Note how the and 
buttons toggle functionality depending
on if the job is currently running or not. When the job is running it cannot be resubmitted, although it may be stopped.
Likewise a stopped job cannot be stopped again, but it may be submitted. If your rendering job timed out without
finishing this can be a good way to finish the render, just be sure to check the skip existing box so that images are not re-rendered.
In this example 7805479.owens-batch.ten.osc.edu is the job id and we can find its output directory simply by clicking
on the the job id.
There will now be a directory like batch_jobs/1567629055 and a file maya-render.7805479 which is the output log for
this job. If you're having issues rendering files, look at these log files for errors. If you need to increase the
log level, see below on how to do that.
Near the bottom, where you should see messages like this. The phrase 'exit status' is an indication on how the job finished and there is help table below to tell you what they mean.
Scene /users/PZS0714/johrstrom/maya/projects/MasterProject/scenes/kai_turntable_02.mb completed.
// Maya exited with status 0
ended at Fri Aug 30 10:00:16 EDT 2019
ended with status 0If you chose to render on more than one node (machine) then you will have several output files like the image below. You'll have one for every node.
This application splits the total number of frames you want to render across all the nodes you want to use (with the remainder going to the last node).
For example if you want to render 20 frames and you want to use 3 nodes, it'll work like this table below:
| Frames | Node |
|---|---|
| 1-6 | 1 |
| 7-12 | 2 |
| 13-20 | 3 |
So if you need to debug frames 9, 10 and 11 you'll need to look at the output file for node to, ending in -2. In the
image above it's maya-render.o7805462-2.
At OSC we run Torque so commonly seen exit status' are documented here for convenience. For more information see Torque documentation
| Exit status | Description & Cause |
|---|---|
| 0 | This is the only good exit status. This means everything went OK, from the jobs' perspective. Maya still could have 'successfully done the wrong thing'. |
| -11 | The job was killed because it ran out of time. Increase the walltime to avoid this. |
| 211 | Maya cannot read your scene file properly. It's either corrupt, or it was never real. Regenerate this scene file in the Maya UI. |
| 271 | The job was killed by a user. |
The default given is -verb -b 1 -ai:lve 0. Here's a breakdown of what those flags mean:
-verbPrint Mel commands before they are executed-b 1By frame (or step) for an animation sequence-ai:lve 0Verbosity level. (0 - Errors, 1 - Warnings + Info, 2 - Debug)
So if you wanted to say, render every 3 frames you would use -b 3 instead of -b 1. Or if you
wanted to turn turn up your log level you could use -ai:lve 1 or even 2.
Note that you can use any combination of arguments. You can remove some or add some independently of each other.
To view all of the possible arguments see this page for the Arnold renderer. Flags for other renderers are not documented here, but could be if requested.
The default for VRay is -verbose=3 which is the default log output level.
See the VRAY help page for all the available flags you can use here.
It's possible for users to run the AutoDesk Maya UI through a virtual deskop application in OpenOndemand (VDI). Users can follow this gist for directions on how to create a desktop icon.
You should already have the ondeand plaform installed and the OSC Repo enabled. If you do, you can simply execute the yum install command.
sudo yum install ondemand-frame-renderer
Create the file /etc/ood/config/apps/frame-renderer/env and add these entries.
# the sftp host your users can connect to to upload and download their project files.
SFTP_HOST=my-host.net
# The cluster your users are going to submit jobs to. We currently only support one
# cluster and this must be a valid entry in `/etc/ood/config/clusters.d/`
#
# defaults to owens (an OSC cluster)
OOD_FRAME_RENDERER_CLUSTER=owens
# The number of cores the job is going to request. This is not variable, and every job
# will request this many cores.
#
# defaults to 28 which, on owens, is an entire node.
OOD_FRAME_RENDERER_CORES=28
# The maya job script that will be templated and submitted. The template language is
# Ruby's ERB. All the variables in the default script will be available to any
# new script file.
#
# defaults to the script template included with this source (relative directory)
OOD_FRAME_RENDERER_MAYA_SCRIPT='jobs/video_jobs/maya_submit.sh.erb'Note, you'll need development eneabled in your ondemand installation for you to develop applications. Or you can check out our images to develop against.
First clone this repo.
scl enable rh-git29 -- git clone https://github.com/OSC/frame-renderer.git frame-renderer
Run scl enable rh-ruby24 -- bin/setup to setup everything up.
Then run scl enable rh-ruby24 -- rake db:migrate to migrate your database.
Now you should be able to connect to <your server>/pun/dev/frame-renderer.
Create a .env.local file if you wish to have environment variables in your development
workspace.