Ignore Patterns and wondering why so slow?

I hope I got the right spot to put this. I did search for about 15-20 minutes on this but “ignore patterm” is quite popular.

My setup: Win10Pro, 2015 Macbook, Android devices, with 2020 Macbook Air M1 soon to be included. Current mentality is WinX and Mac are fully backups, while Android just syncs it’s storage folder and Music folder.

Everything has been working great until today and even did some fairly major file structure changes.

Pattern Matching: I’d like to implement a standard methodology for the entire system. If the file or folder starts with _, it’s to be sync’d. If the file or folder starts with a ‘z’, it’s an ignore.

I’ve implemented (attempted) this as such:

(?d).DS_Store (?d)desktop.ini (?d).thumbnails !_* z*

which when syncing, I saw a ‘zCompost\blah\blah\bling.c’ being sync’d which is not my intention. So then I tried:

(?d).DS_Store (?d)desktop.ini (?d).thumbnails !* !** z* z**

and it didn’t even rescan, just kept on as it had been. I even did a restart and still no change in file count to sync after scan was complete.

So besides the obvs question of “what am I doing wrong” I’d like to ask a pointed question as well. If !_* will correctly handle a root folder (root in the sense of the parent folder for the sync) of “_Sync”, will it:

  1. Sync the entire folder regardless of contents because it has met the condition !_*…
  2. Sync the entire folder but should it encounter a file/folder that starts with ‘z’, then ignore it due to matching the ignore pattern of z*.
  3. Neither. It syncs the _Sync folder and its contents but if there are any subfolders, they will be compared. So if I want _Sync and all subfolders sync, all subfolders need to have the prefix _.

Second question… when I try to edit the Ignore Patterns on the Mac, I get a “Do Not” circle with slash through it but no tool tip or any meaningful indication on what it’s gripe is.

Final question: Experiencing REALLY slow transfer rates… as low as 800Bytes-1.2Kbs, yet speedtest shows full speed. NOTE: All devices on the same SSID. Possible causes:

  1. 80000 files over 6.0Gb… wondering if it’s cuz there’s so many small files.
  2. Is it a Mac to Windows and vice versa thing? I know Mac doesn’t really handle dealing with a high number of low sized files and wonder if that’s what’s happening here.

Thanks. Absolutely love this product. I’ve start degoogling my life and this has been a cornerstone that I’ve been looking forward to tackling. It’s allowed me to build my own Google Drive and almost replace all other device-device software. If I may suggest, sending a link to another device as well as universal copy-paste would make this a single catch all solution. Keep up the great work!!!

The patterns like you wrote above, i.e.


should work fine in my eyes. I would suggest doing a test with a new, clean folder and see what happens there (before even sharing it with anyone else). Do you run the newest version of Syncthing on all the devices?

As for the transfer speeds, I would ignore the slow speeds as long as Syncthing is going through tonnes of tiny files. Syncing such will always be much slower than syncing large files. I would try doing a test with a single large file in order to see what actual maximum speeds you can reach on the connection.

Edit: Just one more thing regarding the slow connection speeds. Make sure that your network connection in Windows is set to “private” and not “public”. Otherwise, Windows will not accept direct connections between the two devices, which will force Syncthing to use relays, and as a result the transfer speeds will suffer. In the older versions of Windows, it used to prompt the user when setting a new connection up, but Windows 10 defaults to “public” and you need to manually change it to “private” in the Settings. You can verify in the Syncthing’s Web UI if the two devices use direct IP addresses to connect with each other, or not.

Thanks for the support.

  • Running latest version except for Android as F-droid is usually a couple days behind. But it’s not involved in this scenario and if off the network.
  • Not only private but all other devices are hard coded in the config.xml with static IPs so Global Discovery, Local Discovery, Relays, and NAT are all shut off. (I don’t care so much for all the chatter they were making (all the way out to the internet!!!) when they are within 5’ of each other.
  • transfer speeds picked right up once the tiny files were done.

Question about the “doing a test with a new clean folder and see what happens”… I’m wondering what I would ‘see’ as it wouldn’t be doing anything other than scanning and building a list to share later no? Please delve a tad further so I know where to look… I’m sure I’ll understand it when I see it.


You can just add a new, empty folder in Syncthing, set your ignore patterns, and then try creating new folders there (i.e. on the disc, not in Syncthing) one by one following your naming scheme. You will see in the GUI whether the newly created folders are ignored or not.


  1. Create a new folder in the GUI with ignore patterns.
  2. Create a new folder on the disc called zTEST.
  3. Check the GUI for the folder status.
  4. Repeat the same process with a new folder called _TEST.

I would doe the same. Make a separate test folder to test your special things and if okay make a transfer to your active folder. Doe you have installed the current v1.15.1 on all your devices?

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