Why does syncthing set a static relay (for example, relay://abc.com), even if the device is directly connected via IP P2P, it still requests DNS resolution every 30 seconds?
And another device has turned off the relay function, but it will still request relay DNS resolution.
What’s interesting is that all listening ports are closed, leaving only the relay address, but it works well and no longer requests for resolution.
I know that when it fails to connect to the static relay server, it will continue to request resolution to ensure that the IP address is correct,But it can actually connect successfully.
Syncthing periodically evaluates available addresses to see if there’s a better one than how it’s currently connected. Apparently things get resolved as part of that. It’s not clear to me how a DNS lookup every 30s is an actual problem, though.
Isn’t IP direct connection not the best way? I read the document and mentioned above that if a direct connection is available, the relay transmission will be stopped.
Syncthing will anyway establish at least one relay connection to have it available, if needed. This doesn’t mean that any data is exchanged via the relay, it just idles around if no connection is currently using a relay.
In this case, let it stay connected (rather than connect → disconnect → request resolution → connect → disconnect → request resolution).
If this is good, why don’t other programs do it?
Maybe when every device has hundreds of programs every 30 seconds, and there are dozens of devices in the local area network, network maintainers will be crazy.
What is even more strange is that even if one of the devices turns off the relay function, it will still request analysis, or that relay switch is useless? This is a bug.