As I dropped my joplin profile folder to sync again from dropbox after having synced from onedrive, I restarted from scratch and now I am going to get everything from dropbox.
But sometimes, I don't know if the process is stuck or if joplin is currently working as no counter move.
So I check the "sync status" and when I go back to the list of notes, the counters move... as if the refreshing of the stats was outdated and refreshed suddenly, or as if i unstucked the process by doing this back and forth.
Finally what could happened ? "i dream" or I spotted something ?. 
Yes it seems to be a glitch in Linux and is just a display issue - the sync is still happening in the background. I think there’s an issue about this on GitHub.
ok, thanks.
Also, I tried to track dropbox networks connection but didnt find.
I had to go back and forth all the evening because the counter didnt move 
If I stay on the status page a long time with thinking the process continues and the counter of objects will continue to increase, is a mistake. It increases just a little bit regarding the 2000 I had to reach.
And once the 2000 I had to wait another 2000 more (but didnt have the strength).
It's very annoying because each time I have to stop the update and restart the day after, I feel that I restart from the ground 0/2000. I could sync almost 1800 objects several times (more than 3days), one after another, but each time I restart with 0/2000.
On the mobile the sync did not take so much time and the count of objects was no so big (I think ~ less than 5000).
On dropbox I just have 4678 files in the Applications/Joplin folder and 3044 others files in Application/joplin/.resource
I'll make a try this evening and hope to see the process reaching quickly the 2000
Yes it does the sync by batches of 2000 items so in your case it would do this 3 times to process your 4678 files. Also if you close the app before it’s completed this initial sync, it will indeed start from the beginning the next time (which I agree is not great), so it’s recommended to let it finish.
That's too bad if before the threshold, everything is lost if i close the app, I think that was better before where this threshold was lower (200 I think instead of 2000), wasn't it ?
i'll try to reach 2000, but yesterday that tooks me from noon to 1 o'clock in the morning to reach 2000
It always worked like this actually and changing the threshold wouldn’t improve things. What it would need is some way to resume the initial sync from where it was instead of from the beginning. When I’ve developed this, I didn’t have many notes so didn’t see this as a problem, but now I can see it’s indeed annoying especially when the cloud storage is so slow.
Is it just Dropbox that is very slow like this or does it do this with all the sync targets? Do you know the total size in MB of your data?
I thought the thresold was like a “commit” with SQL. when reaching a limit you commit to be safe.
With Onedrive I felt that was quicker and I could close the app and open again (maybe i’m wrong with the possibility to open app/close the app/resume the upload where it was, but it’s a feeling) and in 5days the initial import of 2500 evernote notes was imported (total size 650mo).
Since the initial import of all my evernote note, and before you made the support of dropbox, I made a lot of changes, and a lot of deletion, with Joplin and Onedrive.
So now, my profile displays a total size of 180mo used for 2go allowed.
Once the project supported Dropbox, I switched from onedrive to dropbox on the mobile first.
After on my debian I switched to dropbox too, I could finished to sync my dropbox at 1h in the morning after starting the process at 9h30 saturday.
So I don’t know if that was long, on my debian box, because I switched from onedrive to dropbox, but on my win7, after having installed for the first time, It took 1hour for 3500 objects to download and 2400 to be created. It’s not finished but I found that it’s very very fast on windows vs linux.
edit: finished on win7 in 1h30 for:
- 3285 objects locally created
- 5213 downloaded objects
Ok that’s good to know. Not sure what could be affecting the download speed on Linux but if you find out please post here. Maybe it’s an issue with the Electron framework or something else.