microsoft / farmvibes-ai

FarmVibes.AI: Multi-Modal GeoSpatial ML Models for Agriculture and Sustainability

Home Page:https://microsoft.github.io/farmvibes-ai/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

API connection over local network

poysh opened this issue · comments

Dear Farmvibes-Team,

I am trying to access my farmvibes-ai (machine A) installation over the local network from a different machine (machine B), but it seems it is only available on the local machine. Is there any way to open or bridge the API port so I can run the vibes_core client on another machine on the same network?

Hi @poysh,
One approach would be using ssh -L to create a tunnel between machines A and B. Use ssh -L 30000:localhost:3000 <user>@<machine A IP> to create a tunnel. Please, let us know whether it works for you.
Regards,
Roberto

Hi @poysh,
Please ignore my suggestion above and use the approach below.
By default, the cluster endpoint only listens on the localhost address (127.0.0.1). You need to recreate the cluster using the parameter --host to make the cluster listen to all IP addresses available in machine A with --host 0.0.0.0.
The complete command is: farmvibes-ai local setup --host=0.0.0.0
All the best
Roberto

Hey @py5gol,
thank you so much, after some initial setup issues I finally managed to test run the new cluster over the network 🥳
image
One follow up question regarding this. The raster output from the run is on the cluster machine A while the notebook calls the raster_asset.url points to machine B where the notebook is running. Any easy possibity to fix this or should I just put it on a shared drive and grab with based on the run id?

Hi, @poysh. Sorry for the long delay in the reply. I am glad you were able to solve the setup issues and was able to run your workflows.

We currently do not have an easy fix for accessing data from different machines. As you suggested, having a shared drive seems the easiest in your scenario. An alternative would be uploading the assets to a blob storage, from which you could access from the other machine.

Please, let us know if we can help you with anything else.