The two devices that can’t establish direct connections run on OS X.
Thanks for your suggestions. I will test it.
If you run out of ideas with “trial and error”, it might be time to start tracing the network traffic on your LAN and the roter using for example wireshark, tcpdump, dtruss or similar tools in order to locate the paths that the ZT nodes choose to take between the local networks.
Another option is to enable debug logging like in Constant high CPU usage on Raspberry Pi 4 - #6 by zt-joseph
If those are MacBooks you can put them on the same network and see if they connect ZT wise.
Just a quick update. With the static routing between two subnets, now devices under these two subnets can establish direct connection using ZeroTier (albeit redundant). Anyway, it is still weird why they couldn’t without the static routing. Do you mind giving me some pointers on how to use the aforementioned tools for diagnosis?
Yuhong
Glad to hear you finaly had some success!
Since it worked out using static routes over the B network, it’s most likely just a configuration issue in your routers that restrict nat traversal for that partical subnet.
Google “openwrt allow hole punching for zerotier”, “openwrt allow nat traversal for zerotier” etc or ask the OpenWRT forum. Also checkout OpenWRT for NAT examples, reflection and refection zones.
Troubleshooting this type of problem might take some time but when you finally manage to find the root cause it’s usually pretty obvious why it didn’t work (but harder to get there obviously). As for tracing network traffic I ususally prefer WireShark and router logs and try to follow the packet flow from startpoint to endpoint.
Please feel free to report back when you’ve managed to identify the root cause and how you solved it!
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.