Limits
Ignore Crashes
Clients are not deassigned after x crashes, a job is not disabled after x crashes.
No Freeze Detection
The rrClient has functions to realize if a render application has frozen.
If you enable this setting, the client will not abort frozen renderer.
Auto-get Required Memory
The server collects the job memory usage of all clients while they render.
If the memory usage is higher than "Required Memory GB", then "Required Memory GB" is increased.
Max Clients at a time
You can limit the maximum number of clients that should render at the same time for a job.
Note: Do not use this if you have not enough licenses for your render application.
Set the number of render licenses in rrConfig
Required Memory GB
Set the minimum amount of RAM required for the job.
If a client does not have that much available memory, then job is not send to the client.
If a client has multiple job threads running, then the server takes care that the summed "Required Memory GB" will not be higher than the clients available memory.
Default value:
If you have not set any default value yourself, then the rrSubmitter automatically detects the memory usage of the application that submitted the job.
While an application takes more memory while rendering, this is most of the time still a good starting point.
Please see "Multiple Jobs on one computer" for more information.
Max Frame Time
The client is able to abort renderings if the render time is too long.
The value is in minutes per frame.
The value has to be set higher than 1 to have an effect.
Note:
For now, the client is only able to check the time for the whole sequence it got.
For example the client got frame 20-24 to render (5 frames). You have set the max frame time is set to 10 minutes.
If the client takes more than 5*10=50 minutes to render these frames, it will abort it and log an error message. .
Min file size KB
If the rendered file is smaller than this setting, then RR will rename the file and re-render it.
If you have not set any value (=0), then RR calculates a value based on the image verification settings in rrConfig.
If you do not want RR to automatically set this value, then set it to e.g. 1KB for this job or set the size value in rrConfig for the image type you use to 0.
Max Crashes (red log)
If the job crashes this often, the rrServer disables the job.
Max Thresholds
If the number of thresholds reached of all clients are higher than the job setting "Max Thresholds", then the complete job is disabled.
Otherwise a job could be started over and over and is aborted over and over again.
Which is a waste of render time.
Thresholds are reached if ...
- ...a client renders more time than the job setting "Max Render time".
(Useful if you want to prevent that an artist send a non-optimized job by accident. E.g. 2000 frames and a frame time of 20h could block the farm for some time.) - ...the job takes more memory than the machine has.
In this case the clients aborts the job because of its config setting "Abort if system memory gets low".
Restrict Core Count
RR tells the renderer to use less cores.
On Windows, RR sets the core affinity of the render process as well.
Which means no part of the render app can not ignore the core limit setting.
Please see "Multiple Jobs on one computer" for more information.
Restrict Memory
RR instructs the operating system that the process is not allowed to take more memory.
Linux Note:
So far the feature does not work on Linux
Please see "Multiple Jobs on one computer" for more information.
Required GPU Memory
The rrServer does not send a job to a client if it does not have this ammount of free GPU memory.
If e.g. an artist has left open some 3D applications.
Please see "Multiple Jobs on one computer" for more information.
Required Core Count
This feature is mainly used to limit the number of jobs running at the same time on a rrClient.
Please see "Multiple Jobs on one computer" for more information.