search

JuliaComputing / JuliaHub-Feedback

Public repo for filing JuliaHub issues
6 stars 1 forks source link

GPU jobs fail without starting, no error log #129

Open zsteve opened 3 years ago

zsteve commented 3 years ago

Trying to start any kind of GPU instances fails as soon as the job makes it through the queue. This doesn't seem to affect CPU instances.

To reproduce (at least from my account):

Job shows as 'submitted' for a few minutes and then as 'failed'. No other error information displayed from what I can tell. Logs are empty. I've also tried this for Pluto notebook instances and from VSCode plugin, all with the same result.

My account still has the $25 starting amount (as far as I can tell, to be honest the UI isn't the clearest), and I've registered my credit card details, so I'm guessing it's not because of money issues?

Thanks! Stephen

mdpradeep commented 3 years ago

@zsteve , could you please try now ? We had an issue with the GPU nodes earlier today and has since been fixed. Please do let us know if you still have issues.

zsteve commented 3 years ago

@mdpradeep hmm nope still not working. It could be just my account, though :/

zsteve commented 3 years ago

Hmm tried it again just now and it is working. Could this be something related to demand for GPU nodes/insufficient GPU resources causing the request to fail?