This might sound like a stupid question. How is the interval calculated? How is the interval determined?
I can think of 2 possibilities.
1 - The interval is calculated based on the starting of the interval until the interval time has passed. In this way ST does not care if the scan takes a long time or not.
2 - The interval is started only after a full scan is done. This way the internal is a real interval because this way ST actually does not count the scan time at all.
The reason I am asking is that I have somewhat big repos (couple gigs with alot of files) on Arm devices, and I feel like ST is constantly running even If I set them to like 2400s.
Is it possible that the scans take a long time and by the time the scan is finished ST restarts the scan? Some of my repos are on my external sd and encrypted folders and the r/w performance is not super on those.
The set interval is used as the wait time between one scan finishing and the next starting, plus/minus a random fuzz factor.
Well that is fine but what happens when I have 10 shares?
They are independent. You can set individual scan intervals for them.
I know I can set them independently. My question is more about if they overlap? What If I do not want overlap so that my device can rest a while before starting a new scan given that I have multiple shares.
Each folder will run scans on its own schedule, decided by its configured interval, how long the scans actually take, plus a small random offset per scan. The latter is to avoid attempting to scan all folders at the same time, so there’s nothing you can do to force that to happen.
Well, you could disable periodic scans and schedule them via the API, but…
I see. In that case I have to set a huge number than which is not what I want.
I do not know anything about the Rest or the Api stuff. Is there an easy way that I can set this up? I do not mind forcing manual scan on my arm devices. And how do I disable periodic scan?
I’m pretty sure you can disable periodic scans by setting the interval to zero. The do something like
curl -X POST -H X-API-Key:yourapikey http://localhost:8384/rest/db/scan?folder=somefolder when you want to scan a given folder, from cron or so. The rest API docs on the docs site has the details, the above is from memory at the moment (on a phone…).
Thanks for the help. That is good enough for me atm.
Btw can I do this Rest stuff from the browser as well?
If you can configure your browser to send POST requests, sure.
Also, you might be interested in https://github.com/syncthing/syncthing-inotify. This will watch your filesystem for changes and inform them to Syncthing using the same API as described in previous posts. Syncthing-Inotify will also automatically delay Syncthing’s rescan interval as long as it’s running.
Thanks I will see if Inotify works on Arm Linux on Android.
“If you can configure your browser to send POST requests, sure.” Is this possible? I have no idea because.
So I tried the curl method. I am not sure if it is working or now but I get “CSRF Error” at the bottom of the message. At the top I have “301 moved permanently”
Is this expected?
Also is 8384 a rest api port or is this supposed to be the standard 8080 port?
The API key you need to pass with the curl request is missing or invalid.
The rest api port is the same as the web GUI port. In newer Syncthing installs, the default is set to 8384.
Thanks for the help. I will see what the issue is. I thought that I put the right values in place.