How can I let junior node's index be searched at other node?

YaCy YaCy Peer-to-Peer
1

Hello ,I am a developer from China. I try to build a private yacy network. I modified the yacy.init and make a new yacy.network.myname.unit file, in which I change network.unit.name = myname and the network.unit.bootstrap.seedlist0. finally, I build a three nodes network,A ,B ,C. A has a public IP so I make it princple node, B and C are behind NATS. I got a test site on node B , and started a Crawler on B to collect index. after a while, I can do search on node B , but when I search on node A, OR C, I can’t get anything. I have struggled for this issue for many days.
so What can I do so that I can search node B’s content from node A and C?
My original thought was that after B indexes, according to DHT rules, RWI should be distributed, possibly going to node A and node C. So when searching on node A, there should be results, but that’s not the case. Regarding this issue, I have some guesses:

  1. B’s indexing information was not transmitted by DHT but saved locally on node B. This might be because the DHT distributed algorithm calculated that this website content should be stored locally on node B.
  2. B’s index was transmitted, but when initiating a query on A, the query request was located to node B by DHT. However, since node B is behind NAT, it cannot be directly accessed, resulting in no results being found. But shouldn’t the query request check local index entries first? This is contradictory.
    So, I’m a bit confused here. Please help.
2

It seems like A cannot access B and C, because they’re behind NAT?
B and C should be able to send RWI index to A, but the process is rather slow and random. (bothers me as well, I’d prefer faster/larger RWI distribution).

You can use some sort of SSH tunnel or VPN, to keep them on one network, at least internally.

If on the same network, you can also use the index of “indexing” nodes as a remote solr.

Or, you can setup a separate solr or solr cloud server to serve all 3 instances. In that case, it’s IP address has to be accessible from all the peers as well. See Index Administration > Index Sources & Targets.

1 Like
3

Hi @yacy_wxb

Have you asked ChatGPT 5.2 for a solution?

I quired your exact post and then “exact yacy.init settings for this topology”.

ChatGPT suggests having 3 configurations.

I am attempting to “Reverse Tunnel - Matching Public IP Address to Actual IP Address”

You are welcome to try this copy of Yacy https://github.com/smokingwheels/yacy_smoke

4

thank you for your replies, I give up this way and change my to another plan to achive my project.
I think maybe,Yacy consider B and C are behind nat,so there is not mean to let the tow nodes’s index tobe transfer arcoss the network…

5

thank you for your replay. I have tried this action,and no change happen.
I want to use yacy to build a network which contain many nodes behind nats.
I used to consider this project can make them search each other’s data,but…

6

maybe the VPN to glue all three nodes together? with a private addresses and use them as a external solr address in the public instance?