Destination Host Unreachable From Node

[Lesther Santana Carmona]

To test network connectivity on a newly created node, I used “ping 8.8.8.8” and got as reply “Destination Host Unreachable.”

Also, based on basic_gpu_devices.ipynb I tried to add the Cuda repo but got “Failed to connect to developer.download.nvidia.com port 443: No route to host”.

What might be causing this issue?

[Devin Lane]

Can you try doing ‘ping google.com’? I believe I had a similar issue with the IP not resolving, but specifying the domain name worked for me.

[Lesther Santana Carmona]

Yes, ‘ping google.com’ works. But how can we solve the “Failed to connect to developer.download.nvidia.com port 443: No route to host”  issue?

The command that I am trying is:

sudo dnf config-manager –add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel8/x86_64/cuda-rhel8.repo

[Paul Ruth]

Which FABRIC site is your node on?

[Lesther Santana Carmona]

It’s in CLEM.

Image: default_rocky_8
Image Type: qcow2
Host: clem-w1.fabric-testbed.net
Site: CLEM

[Paul Ruth]

This is because that Nvidia site is not setup for IPv6 and FABRIC’s CLEM site uses IPv6 addresses on its management network.

One reasonable workaround for this is to use the NAT64 services described here: https://nat64.net/

There is an example FABRIC Jupyter notebook in your JupyterHub container called  “Accessing IPv4 Sites from IPv6 Nodes” that shows how to set this up on a FABRIC node.

Paul

[Lesther Santana Carmona]

Thanks a lot for the information!

[Author]

Posts