Why can't 2 machines running ipfs daemons on same local net share files?

Help Kubo
1

I’ve been working with ipfs for 4 years now but I’m totally puzzled by this problem, I’ve not seen this before. I must be missing something obvious.

These are desktop Linux machines, both on same 192.168.2.x subnet. Both have ipfs companion browser extension, and both have over 1000 peer connections which change dynamically. They both can access the webui.

but they can’t access each other!

I added a peer definition to each of the server configs in the “Peering/Peers” section for the other node like so:

  {
    "Addrs": [
      "/ip4/192.168.2.180/tcp/4001"
    ],
    "ID": "12D3KooWDsudkyLhzdMVG6Rw2qvnKEPKCEeGuqLky2My3wNSq32m"
  }

But it doesn’t help. I can add files to each server and get them back on the same server, no problem. When I attempt to get a file on the other system it hangs & doesn’t return.

Anyone have suggestions on this? I’ve never experienced such an issue and it’s driving me crazy.

1 Like
2

Does the other local peer show up in ipfs swarm peers? Additionally, does it work as expected if you include the peerID in your Addrs array? IE: /ip4/192.168.2.180/tcp/4001/p2p/peerID. Finally, does ipfs swarm connect <multiaddr> yield any information when attempting to connect manually?

3

Would you, by any chance, have anything defined in:

"Swarm": {
  "AddrFilters": [
    ...
  ],
1 Like
4

After checking the two messages above.
Can you show ipfs stats dht lan after the two machines has been running for ~10 minutes ?
If they have failed to discover each other that means your network is about as good as WAN connections.

1 Like