Split tunnelling versus IP filtering?

In order to set up a shared folder between my PC and raspberry Pi, I need to set up split tunnelling for my VPN. This requires having an .exe or something similar to attach chia to. Has anyone had success with connecting Chia to split tunnelling?

I am using PIA VPN. Instead of using the app-based split tunneling, I filtered the IP that I want it to bypass the VPN.

In fact, in the PIA client, you can allow LAN connection without having to manually filter your LAN IP.

If you don’t mind, could you help explain IP filtering and VPNs for chia? For example, I am plotting on my PC, but want to transfer my finished plots to a harvester on a Pi. Both machines are connected on my LAN, but my VPN (Ivacy) allows only one device to be connected to the dedicated IP at a time.
I’m assuming that I will run the Pi on this dedicated IP. Would the IP filtering take place on the Pi, or the PC?

Why do you need a VPN if both machines are on the same Local Area Network?

The havester needs to connect to the farmer node directly through the local IP on port 8447 by default. You don’t need a dedicated IP on the harvester.

The first thing you need to know is your LAN IP. Then, filtered the IP out of your VPN connection. My LAN IP is The /24 means the last 0 can be any device in my LAN.

I don’t actually know what I’m doing. Both machines are connected to my router through ethernet. I clearly need to look up what a LAN is.

Ok, so if I don’t need a dedicated IP to harvest, then the Pi doesn’t need a dedicated IP or even to be connected to a VPN? And a LAN may not even be necessary if both machines are independent? I have had a lot of difficulties connecting to 8447 and 8444 in the past because I’m behind a CGNAT.

What is it that you’re actually trying to accomplish?

Create a shared folder accessible over the network?
So that you can drag plots from your plotter (windows) to your harvester (pi)?

Yes, I want to create a shared folder. I also want to be able to run the harvester independently on the Pi but have had so much trouble without a dedicated IP I’m not clear if I’ll need to run the PI with a dedicated IP or not.

Ha. You’ve lost me. So you want to run Chia Farmer on the Pi? i.e. to farm your plots independently from any other machine? But having problems?

Just to be clear, you need advice on:

  1. How to set up your Pi as an independent farmer.
  2. Create a folder on your Pi to drag and drop plots into over the network.
  3. Do the above behind a CGNAT.

The concept of the harvester is that you only have to run a single farmer node (the one that syncs to the full node) while the rest in your network run on the harvester nodes without having to sync to the full node.

Therefore you don’t need any VPN, let alone the dedicated IP, on the harvester unless you want to remote control your harvester over the internet.

And it’s true that you would have a hard time opening a port if you’re behind the CGNAT. However, all you need on your harvester is the local connection. Your local connection/LAN has nothing to do with CGNAT. You should look at the firewall of both the farmer and harvester, and make sure that port 8447 from LAN IP is allowed on the farmer node machine so that the harvester can connect.

I can’t recommend this method, especially if you are running the harvester on the Pi. I am assuming that you use Samba, which is super slow, for this.

What I would recommend is that you set your plotter as a harvester and your Pi as a farmer (full node). This way, as soon as your plotter finished a plot, it will be farmed automatically. And you can swap HDDs around. But don’t use a network transfer to transfer your plot. On the Pi you are most likely getting around 10 Mbps with that :sweat_smile:

The shared folder is to pass the db, ca and wallet folders between machines, not the plots or any other data. I intend to only run the Pi, so that I can disconnect my PC from Chia.

1 Like

Yes, this is good. You can create a shared folder without having to worry about the VPN or split tunneling.

However, if you want to run Chia entirely on the Pi, you will have to run it as a farmer, not a harvester.

Here’s my issues as best as I can describe them:

  1. The Pi was previously running as an independent farmer for OG plots. I would like to continue to run the Pi as an independent pool farmer, but am trying to understand if I need to disconnect my PC from the dedicated IP or the entire VPN in order for that to work (I’m behind a CGNAT and my VPN only allows one device for the dedicated IP).
  2. The shared folder is to drag and drop the ca, db and wallet files
  3. The CGNAT seems to be preventing pool farming, as well as many 8444/8447 connections with OG plots

And you’re running the Pi as a lightweight desktop computer with a monitor, mouse, and keyboard? (As opposed to a headless remote server with no monitor).

That’s helpful. I believe that’s what I had done before.

It’s headless. Previously I had checked in on a mobile ssh app, and done folder transfers using Remote Desktop.

I cant get to my PC atm, but to have a folder accessible over the network you’ll need to learn a little about samba.

Either set up a samba server on your pi and share a folder to the network, then add that folder as a remote drive on windows.

Or, share a folder in windows and figure out how to mount it on the pi with the samba protocol.

If no one has solved your issue then i’ll add a post on how to get set up when i get home. :+1:

Unfortunately, I’ve been unable to connect through sambashare because of the VPN/IP issues.