Is there an expiry on Syncthing installations / IDs?

#1

I have 2 separate Synthings running on VMs for different clients. I act as a receive only offsite backup. Couple of days ago I turned off one of the Synthings in order to pull down files that refused to sync via FTP. Now, when I fire that instance up some 2 days later all 9 remote offices are showing disconnected and have done for the last 8 hours.

The other Syncthing is still happily syncing away which suggests comms are good.

I’ve restarted Syncthing, computers, turned off firewalls, moved the St folder to another machine, even had a go at renaming and forcing a new db, but no show. The end points that ‘send only’ have both my IP:22000 (port fwded here at the remote backup) + dynamic. I turned off the good St but the failed one didn’t do anything so thats back up and running and is fine.

Only thought I have is maybe the discovery server has expired my IDs and is why it’s no longer seeing anything. If this is the case it suggests that peoples Syncthings can’t be off for long.

I can wipe away the current setup, but it’s a big task to remote into the offices and setup all the jobs.

Is there a test I can do to check the ID or connection is valid?

(Jakob Borg) #2

There’s no expiry of the kind you’re thinking of.

Discovery announcements are expired and removed after an hour, but a device that is online will refresh it every 30 minutes.

If you expand one of the disconnected devices you can see if it’s found any addresses for them. In an upcoming release, you’ll be able to see why it can’t connect (if that’s the problem).

#3

There’s no connection / ip / relays on any of them

image

(Jakob Borg) #4

Then they are not online, or not announcing themselves to global discovery.

#5

Is there a way to check, a command line or similar?

there seems to be some sort of chatter going on… image

on a restart it seems to go to the discovery server then goes to sleep

#6

Im up and running. In the end I unticked global discovery, restarted, then reticked it, restarted on all the devices and they eventually connected.

Don’t know if that corrected anything in the code / config, but seems odd that all 10 devices had to be done in order for them to reconnect.

Mentioned here incase it helps anyone else in the future.