python - Flexible task distribution in workers with Celery -
TL; DR: Is there a possibility to easily fix workers their (local) configuration and task args
/ kwargs
The idea of a quick and dirty solution will be in all those employees who find themselves inappropriate, but I was hoping someone else Is elegant.
Description:
The application is an (educational) programming assignment assessment tool - possibly comparable with continuous integration: a web application (previously Specified) Accepts submission of source code for programming languages (or better: programming environments), which then needs to be compiled and executed with multiple tests. Now to use in high performance computing courses especially with GPUs, the compilation and execution can not be on the host where the web application runs (for other reasons it seems only for security reasons).
Administrators, I need a configuration file for a worker to configure it easily, where locally available resources, compiler types and paths are configured, for the worker to decide Uses whether or not to work on the job.
Just by using different queues and using custom, I do not appeal, because the number of queues and configurations may vary in runtime and I may feel a little suspicious.
Is there any great way to achieve something like this? To be honest, the documentation did not give me much guidance on this.
Thanks in advance for any suggestions and hints.
Comments
Post a Comment