Syncthing 18.x and then 19.0 stops syncing when primary name server is down

I run local caching name servers here. one a pine64-LTS SBC with Arabian linux. The other is a VM on one of my servers running Debian 11.

I shut down the primary to move it’s location and got distracted reconnecting it and power. the secondary was still up and handling dns lookups just fine for my local network clients.

Noted that my Syncthing instance stopped even connecting to my remote. Log message was limited to a notice that no copy of file X was available on any devices.

spent some time troubleshooting. I had connectivity, could look up the domain name of my remote machine ping it, log into it. No issues.

Remote is an ubuntu server 20.04 VPS, local is a Debian 11 VM. at the time running the last v18 release of Syncthing. now 19.0

I decided to take a break, clear my head and went to plug in and connect the primary name server to the network. When I came back to troubleshooting, it was syncing again.

powered the primary NS down again and it stopped syncing…

edited my resolv.conf file on the local VM that runs Syncthing to list my secondary NS first, restarted Syncthing on it, and it started right up syncing again.

I can reliably repeat this behavior.

I do not need help, just wanted to pass on this glitch I’ve discovered.

The Go runtime uses its own DNS resolver, perhaps it parses your resolv.conf differently from your system resolver. Or just works differently when the primary resolver is down.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.