Worker Machine Went Offline?

[ryan_otterson]

Hello,

My recent render, sent twice to the cloud at appx 8000px by 4000px failed twice saying "something went wrong" and "the worker machine went offline"


The job id was KILaZHDrFT7BMet4. Im not sure why it would fail on a render size such as this. Can you help me?

[vladie]

Hi ryan_otterson,

I'm sorry for the slow reply. I believe my colleagues have replied to you in the Support ticket you have opened.

Just to summarize the resolution for the rest of the users here. The issues seem to be connected with memory issues with the GPU cards on which the jobs are rendered. The 16 GB of the T4 cards that we use for V-Ray GPU jobs seem not to be enough for these jobs. So our advice is to reduce memory consumption of the projects. You can check these guides for reducing the memory consumption:

1. Memory Usage Optimizations for GPU rendering
2. Memory Usage Optimizations for CPU rendering
3. V-Ray Displacement Optimizations

Alternatively you can submit a V-Ray GPU job using the CUDA CPU Override option in Chaos Cloud which will render the job only on CPUs in CUDA mode. This way it will benefit from the 128 GB of RAM that the CPU machines currently have.

Please let me know if you try any of the workarounds and how it goes.

Thanks,
Vladie