It takes long time to load the Jupyter webpage on an offline server from another offline machine

[Jason-Hsu]

Describe the issue:

I deployed the ml-workspace container on a RHEL server, then use another machine to connect to this container. Our environment is an offline environment (both the server and the client). It takes 20~30s to open the page. Also, even if I log in to use it, it sometimes takes long time to load page after using it for more than one hour. However, if the server connects to the internet, then the webpage's response time becomes few seconds.

Technical details:

• Workspace version : 0.13.2
• Docker version : 20.10.9 (also test on 20.10.16)
• Host Machine OS (Windows/Linux/Mac): RHEL 8.4 (also test on Windows 10 machine as the server)
• Command used to start the workspace : docker run -p 8888:8080 mltooling/ml-workspace:0.13.2
• Browser (Chrome/Firefox/Safari): Chrome and Firefox

[Jason-Hsu]

Finally, I found a solution to solve it. It's related to the nameserver setting. After I remove the nameserver setting in /etc/resolve.conf, the issue is gone.

[github-actions[bot]]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 14 days