Need help configuring private swarm

Help
1

Hi, I’ve been trying for a few days now to setup a private network of ipfs using swarm.key so they can all be in a private network.

I’ve set up a node on my local machine, another node dockerized on my local machine and a third node on a cloud service ( also dockerized ).

I’ve confirmed the ports are open and I can ping each node from inside the others. However if I try to netcat into the port 4001 ( the one that handles the swarm connections ) I get weird characters that I can not put in here because the post will get removed for some weird reason, I will share a screenshot instead.

image

On the other hand, when configuring all the nodes. I get the correct intialization using the swarm.key and they are all on private mode. I remove all bootstrap nodes and start from 0 to configure manually the connections. After adding the node 0 to every other node I then try to connect using ipfs swarm connect, it does not matter from where I try to connect I just can’t. I get the following errors:

Error: connect failure: failed to dial: failed to dial : no good addresses

  • [/ip4//tcp/4001] gater disallows connection to peer
  • [/ip4//tcp/4001] gater disallows connection to peer

Some help would be much apreciated.

Thanks!

2

Do you have addrFilters set in the config and is the peer’s IP in one of them? I.e. private ranges when initialized with server profile.

3

I do have addrFilters set, but I did not set them manually, should I have ?
And no, the ip add of the other containers are not in the main one or on the others

4

I have removed all addrFilters from both nodes ( even though the other nodes were not getting filtered, or at least I thought so )

Now I’m getting the following error when trying to connect to one of the nodes.

Error: connect failure: failed to dial: failed to dial : all dials failed

  • [/ip4//tcp/4001] failed to negotiate security protocol: incoming message was too large
  • [/ip4//tcp/4001] dial tcp4 0.0.0.0:4001->10.99.2.2:4001: i/o timeout

Also, should the node0 be initialized as server mode? When I tried to connect the nodes in non-private mode I just had them all as regular nodes.

I’ve also set the nodes as debug mode so I can see the logs, when trying to connect from one node. The other one spits out this log:

2024-05-09T05:59:48.024Z DEBUG upgrader upgrader/listener.go:126 accept upgrade error: failed to negotiate security protocol: incoming message was too large (/ip4//tcp/4001 ↔ /ip4//tcp/57156)

5

I finally managed to solve the issue.

I set both my local node and the external ones as dhtservers, then restarted the whole process from scratch just to be sure everything was correct. I got the nodes connected properly!

6

I think I’ve seen this when the private key is different between nodes.