Syncthing clients stopped synchronization

I was testing syncthing how it will function in a larger client environment and I had a finding. I tested syncthing with 48 clients and 50 folders with different size (total of 20 GB). At first the synchronization went fine. But after a hour or so I noticed 5 clients stopped the synchronization.

I checked the following:

  • The syncthing process is still running
  • The rest API is still available.
  • Calling GET /rest/system/config works fine
  • Calling GET /rest/stats/folder does not work. It results in a timeout. Other calls involving the database, like statistics, also don’t work.

After restarting syncthing, or killing the process and restarting, the synchronization continues and completed successfully.

We are using syncthing version: 0.13.2.

I did not notice anything in the syncthing logfiles, with debug level ‘events’ (POST rest/system/debug?enable=events). I also checked the windows eventlog. There is nothing there related to syncthing.

Can you help me to investigate this issues. How can I approach this to generate enough information to solve this issues? Or is this a known issue/limitation of syncthing?

If one of the failing nodes is running on linuX, you can send a sigstop to get the stacktraces.

Otherwise check -help on how to enable pprof, and then provide content of /debug/pprof/goroutine?debug=2 call.

SIGQUIT it is. SIGSTOP will just pause it.

Yes, thats what I meant :frowning:

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