Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve docs for deployment via Azure VM #491

Open
6 of 7 tasks
ncclementi opened this issue Jan 10, 2025 · 1 comment · May be fixed by #499
Open
6 of 7 tasks

Improve docs for deployment via Azure VM #491

ncclementi opened this issue Jan 10, 2025 · 1 comment · May be fixed by #499
Assignees

Comments

@ncclementi
Copy link
Contributor

ncclementi commented Jan 10, 2025

The instructions to create a VM via Azure in here https://docs.rapids.ai/deployment/stable/cloud/azure/azure-vm/ are not very clear if this is the first time someone is creating a VM.

I went through the process and I made/guess some choices but we can be more diligent so people can fill in the blanks.

To improve:

  • it's required to create a resource group as part of the VM creation. Mentioned or link to what this is or what people should put
  • We suggest that people use the NVIDIA GPU-Optimized VMI when selecting that one, it prompt to me two options that ahve the same version but one said Gen 1 and the other Gen 2. I wasn't sure which one to choose or if it mattered.
  • By default the Security typed is set to Trusted launch virtual machines but this causes an error with the choice of NVIDIA GPU-Optimized VMI. Then I chose Standard security type to carry on, but I wasn't sure if this was correct.

Image

  • We have a recommendation of VM sizes but it's not visible in or linked in the document you are following to the process. We should link to https://docs.rapids.ai/deployment/stable/cloud/azure/
  • We should also suggest that people create the ssh keys or mentioned something about that section.
  • When creating the security group, I opted to not modify anything except the destination ports as suggested, but I wasn't sure if that was the right move.
  • Finally, when connecting to the VM, we point to the Azure connect instructions, which are very easy to follow. One caveat is that when you try to follow that it fails on ssh because the .pem gets downloaded with more permissions and the connection fails. This can be resolved by doing a chmod 600 the_key_file.pem
@ncclementi ncclementi self-assigned this Jan 10, 2025
@jacobtomlinson
Copy link
Member

A few thougts:

  • Folks often have an existing resource group for their project. They probably don't need to create one for each VM.
  • I assume we want to use the Gen 2 images.
  • We may want to feed back to the team that manages the VMI about the security warning. For now we should just document that it needs to be changed to standard.
  • For port access all you need to do is ensure the reader can get from their local machine to Jupyter and Dask. We shoudl document the minimum required to do that.
  • Shame the PEM is too permissive, we should document the chmod too if it's necessary to get things working.

@ncclementi ncclementi linked a pull request Jan 24, 2025 that will close this issue
1 task
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants