Bryce Ferenczi

Inserting lines at torchx/schedulers/kubernetes_scheduler.py:ln343, just before status = resp['status'] ```python3 for key_, obj_ in resp.items(): print(f"{key_}: {obj_}") ``` Output is below: apiVersion: batch.volcano.sh/v1alpha1 kind: Job metadata: {'creationTimestamp': '2021-08-07T06:56:36Z', 'generateName': 'echo-',...

Yes, I was just in the process of commenting that it seems like things are running, but as soon I peeked behind the curtain at the logs I saw that...

Unless there is a more proper way to check to see if the job has launched successfully, I think this could be treated as an indirect measurement of determining that...

No worries for the late reply, my patch is working so nothing is urgent, but not for the reason I expected, which is why I created this issue. Both queues...

Certainly, but I think maybe better intergration of the existing nodeSelector mechanisms is cleaner than another plugin and set of attributes, which is what I'm using *mostly* successfully.

Before this intergrated, I just wanted to bring up the point of why not just properly intergrate with nvidia's gpu operator? I made a fork of volcano and made modifications...

> Pls help take a review. Thanks! If possible, I plan to put this design into this weekly meeting for a statement. @william-wang @shinytang6 @k82cn If this is during outside...

Just a brief update, nvidia now have an [oversubscription model](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/gpu-sharing.html) in their gpu-operator that could be intergrated into this.

Do you have an example library where you run stub generation during the building process (pip install .)? I have this [project](https://github.com/5had3z/sc2-serializer), and I would like to somehow get the...

> @5had3z see [#104 (comment)](https://github.com/sizmailov/pybind11-stubgen/issues/104#issuecomment-1706128213) I'm able to get it in a single pass [here](https://github.com/5had3z/sc2-serializer/blob/main/setup.py) with my additional flag.