macOS Tahoe 26.6 + Desktop 34.0.1: client connected but local changes are never detected after update

Hi everyone,

I’m experiencing a synchronization issue that started immediately after upgrading the Nextcloud Desktop client to version 34.x.

Environment

  • macOS Tahoe 26.6
  • Nextcloud Desktop 34.0.1
  • Hosted Nextcloud server (Aruba)
  • External drive used for local sync
  • Dataset: approximately 1 TB

Problem

The client connects successfully and always reports:

“Everything synchronized”

However, it never detects any local file changes anymore.

If I create, modify or delete files inside the synchronized folder, absolutely no synchronization activity appears.

The remote server is fully accessible and contains all files up to about one week ago (before the upgrade).

What I’ve already tried

  • Updated macOS to Tahoe 26.6
  • Updated Nextcloud Desktop to 34.0.1
  • Restarted the Mac
  • Reset File Provider Domain
  • Verified the account is connected
  • Verified server accessibility
  • Verified remote data integrity

None of these changed the behavior.

Additional information:

The local files are intact.
The remote files are intact.
Only new local changes are ignored.

Has anyone seen this behavior after upgrading to 34.x with Apple’s File Provider?

I would like to avoid removing and recreating the account because the dataset is over 1 TB.

Hello, @Marisa!

For now, I can only tell that we test the synchronization of remote and local changes in two different scenarios: when a server does have a high performance backend installed and enabled (“push notifications” for file changes) and when it does not. In our test scenarios, this works fine.

I assume the devil is in the details of your use case. Since you already mentioned a data volume of around 1 TB: How is it composed? Do you have mostly large files like videos or similar types which usually are large? Or do you have a situation with rather thousands, hundred thousands or higher orders of magnitude in terms of file quantity? A recent topic we are confronted with are folder with thousands or more files which can slow down Nextcloud servers (depends on the infrastructure used and resource available) so much that the invisible background activity of fetching the news from the server takes so long that the client appears stuck with what it has.

Hi Iva,

Thank you for your reply.

Regarding the data set, it is my archive of about 1 TB stored on an external Toshiba drive. It mainly contains:

  • Adobe Illustrator (.ai) files
  • Photoshop (.psd) files
  • PDF documents
  • Images / Fotos (.jpg, .png)
  • Music
  • Office documents
  • Fonts
  • Other design assets

There are many folders and subfolders, but this exact dataset synchronized correctly until I upgraded to Desktop Client 34.x.

After doing some additional testing, I noticed something that may help narrow down the issue:

  • Server → Client synchronization works correctly.
    If I upload a file or create a folder through the Aruba Drive web interface, the Desktop Client (Nextcloud) immediately detects the change, downloads it locally, and shows it in the Activities list.
  • Client → Server synchronization does not work.
    If I create a new file or folder inside my local synchronized folder, nothing happens:
    • no activity appears in the Desktop Client (Nextcloud);
    • the item is never uploaded to the server;
    • pressing “Sync now” does not trigger any visible synchronization.

I also noticed that the Finder context menu no longer shows any Nextcloud actions (such as keeping files downloaded or freeing space) inside the synchronized folder.

The synchronized folder is located on an external Toshiba drive, and this exact setup worked perfectly before upgrading to Desktop Client 34.x.

One more observation: the issue started immediately after upgrading to Desktop Client 34.x. No changes were made to the server, the folder structure, or the external drive itself.

Could this indicate a problem with detecting local filesystem events or with the macOS File Provider integration rather than the dataset size?

Thank you!