Assuming you have bare metal machines, yes. You can also install it on top of K8s. Are you using bare metal machines on prem?

I'm using at the moment 2 GCP instance:

1. for clearm-server
2. for gpu-cluster (A100 with 8 GPUs)

In that case then yes, install the agent on top of the machine with the A100 with 8 GPUs

the agent should be installed in the remote resource right?

many thanks!!

it shouldn't happen through the resource configurations tab?

how?

This is doen via clearml-agent this is the second link I provided - None

None

No, resource configuration allows you higher level of control over different resources - None

I suggest you review the agent setup & deployment docs - None

Then you can view all the connected nodes through the orchestration tab - None

In the OS you can create queues and manage your clusters through that

thanks @<1523701070390366208:profile|CostlyOstrich36> ! so can you refer me to the documentation the describe how to connect my GPU cluster into the clearML-server?

so in the open source version I can't create queues, manage all GPU clusters?

Happy to help 🙂

I suggest watching this short video to get a better idea on how the agent works - None

Hi @<1798162812862730240:profile|PreciousCentipede43> , role based access controls, admins and resource configurations are all part of the Scale/Enterprise licenses. In the open source everyone is set as an admin with access to all parts of the system.

Hope this helps 🙂