Can I use EDGE=1 for Syncthing?

[JustinChase]

I have Syncthing installed, but when I run it, it shows an update is available.  I understand the EDGE variable is supposed to allow us to update inside the container, but I also think it must be 'enabled' in the container/docker first.

 

So, before I add EDGE = 1 to my template, I thought I should check to see if that should work with the gfjardim/syncthing docker.

 

thanks.

[gfjardim]

I have Syncthing installed, but when I run it, it shows an update is available.  I understand the EDGE variable is supposed to allow us to update inside the container, but I also think it must be 'enabled' in the container/docker first.

 

So, before I add EDGE = 1 to my template, I thought I should check to see if that should work with the gfjardim/syncthing docker.

 

thanks.

 

Syncthing has a built in auto update function, so no EDGE routine is available.  Do you have any problems doing an in app update?

[JustinChase]

I haven't tried the in-app update yet.  I thought doing so could be bad, since reinstalling the container at a later date from the docker would install the version I currently have, not the version I'm about to update to.

 

I thought EDGE was the way to handle this.

 

I could be misunderstanding the whole thing though, so I defer to you for how to best proceed.

 

thanks.

[cirkator]

Sorry to bring up this old thread but i think the update routine of syncthing has changed. I remember being shown an icon while beeing on the syncthing page that an update is available. Clicking this icon updated the docker just fine.

Now i dont get that icon anymore and therfore cant update syncthing which is running v.0.9.18. The current version is v.0.10.5.

 

Can anyone confirm this issue?

Solutions?

[cirkator]

I had a problem and needed to start with a fresh docker image. Upon reinstalling the syncthing docker fresh, i got the in app update function back. Just updated succesfully to the newest version. All good

[boohbah]

I have playing for days with gfjardim's syncthing Docker and here is what I have learnt.

 

If network type is set to 'Bridge', then the online version update function will work, but the local discovery function will not (regardless of your port routing skills). So if you want to connect and sync between nodes on the same LAN, you can't. If you set your local nodes to use online discovery, then announce.syncthing.net gets confused about who it is talking to after the NAT in your router. This causes weirdness.

 

If network type is set to 'Host' (with necessary ports 21025, 8080 and 22000 all specified), then the online version update function will not work (you are forever stuck at 0.9.18), but the local discovery function does work - and very nicely. In this condition, connection and sync works between same LAN nodes just fine.

 

So, at present, we can't have it both ways.

 

Ideally, at least for me, I would like to be able to have 3 same LAN nodes all sync together, and have just one of those connecting with the announce server for off site nodes (later on). Of course the cherry on top would be to have them all able to auto update their version. Version 0.10.12 came out about 8 hours ago 

 

I don't know how to fix this. Perhaps the maintainer can take a look?

 

[gfjardim]

I have playing for days with gfjardim's syncthing Docker and here is what I have learnt.

 

If network type is set to 'Bridge', then the online version update function will work, but the local discovery function will not (regardless of your port routing skills). So if you want to connect and sync between nodes on the same LAN, you can't. If you set your local nodes to use online discovery, then announce.syncthing.net gets confused about who it is talking to after the NAT in your router. This causes weirdness.

 

If network type is set to 'Host' (with necessary ports 21025, 8080 and 22000 all specified), then the online version update function will not work (you are forever stuck at 0.9.18), but the local discovery function does work - and very nicely. In this condition, connection and sync works between same LAN nodes just fine.

 

So, at present, we can't have it both ways.

 

Ideally, at least for me, I would like to be able to have 3 same LAN nodes all sync together, and have just one of those connecting with the announce server for off site nodes (later on). Of course the cherry on top would be to have them all able to auto update their version. Version 0.10.12 came out about 8 hours ago 

 

I don't know how to fix this. Perhaps the maintainer can take a look?

 

I've just tested here, and no problems with the self-updater on bridge/host network. When bridge network is used, no peer auto discovery/UPnP is available because this kind of stuff (mDNS/UPnP/Bonjour/Avahi) only broadcast on the local subnet. You can always remedy this by using some sort of dynamic DNS (freends.org is awesome) and port forward set on your router, and then add the unRAID client on your other clients with the address you setup.

 

 

[boohbah]

Good ideas.

 

I have also just discovered that in 'Bridge', I can get the local LAN nodes to recognise each other if I manually enter the static IPs of each into the respective address boxes on each (rather than leave it as 'dynamic'). This works with the IP address alone, and no port needs to be specified (which the gui invites you to do).

 

So. I guess that could be called a fix.

 

 

I'm very grateful for this docker by the way. Thank you so much.