Let’s see what we can find:
% cd ~/Library/Logs/kopia/cli-logs
% ls -lh
-rw-r--r--@ 1 user staff 291B Nov 16 15:19 CACHEDIR.TAG
-rw-r--r--@ 1 user staff 48M Nov 17 22:00 kopia-20251116-141902-18634-server-start.0.log
-rw-r--r--@ 1 user staff 48M Nov 18 08:00 kopia-20251116-141902-18634-server-start.1.log
-rw-r--r--@ 1 user staff 48M Nov 18 18:00 kopia-20251116-141902-18634-server-start.2.log
...
Right… The log files are HUGE! And deep in there are clues hidden to solve my puzzle? 
Let’s see what we can “grep”:
% less kopia-20251116-141902-18634-server-start.2.log
2025-11-18T07:19:33.022890Z DEBUG scheduler triggering refresh
2025-11-18T07:19:33.025665Z ERROR connection connection closed: unrecognized error when opening SFTP file macmini/kopia.repository.f: connection lost, will retry
2025-11-18T07:19:33.025738Z DEBUG connection closing active connection.
2025-11-18T07:19:33.025782Z ERROR connection error closing active connection: error closing SFTP client: EOF
2025-11-18T07:19:33.025821Z DEBUG retry got error unrecognized error when opening SFTP file macmini/kopia.repository.f: connection lost when GetBlobFromPath (#0), sleeping for 100ms before retrying
2025-11-18T07:19:33.127900Z DEBUG connection establishing new connection...
2025-11-18T07:19:33.128062Z DEBUG sftp using internal SSH client
2025-11-18T07:19:35.719193Z DEBUG scheduler >>>> scheduling 3 items (now 2025-11-18T08:19:33+01:00)
2025-11-18T07:19:35.719358Z DEBUG scheduler 2025-11-18T12:19:33+01:00 refresh
2025-11-18T07:19:35.719404Z DEBUG scheduler 2025-11-18T08:42:11+01:00 maintenance
2025-11-18T07:19:35.719433Z DEBUG scheduler 2025-11-18T09:00:00+01:00 snapshot "/Users/yoriz"
2025-11-18T07:19:35.719459Z DEBUG scheduler sleeping until 2025-11-18T08:42:11+01:00 (maintenance)
...
So there are connection errors apparently? Interesting. Never got these with Duplicati so I wonder whether there are really connection problems, or incompatibilities with the behavior of my SFTP provider and Kopia. However, it seems to be able to (re-)connect after retrying.
What does this mean? Was the backup successful? Was the backup aborted?
The next error is more clear:
2025-11-18T08:00:02.335255Z DEBUG uploader snapshotted directory {“path”:“Pictures/Photos Remote.photoslibrary”,“error”:“cannot create iterator: unable to read directory: open /Users/user/Pictures/Photos Remote.photoslibrary: operation not permitted”,“dur”:“66.584µs”}
2025-11-18T08:00:02.336413Z DEBUG uploader snapshotted directory {“path”:“Pictures”,“dur”:“1.554041ms”,“size”:9535,“files”:2,“dirs”:6,“errors”:1}
2025-11-18T08:00:11.322698Z DEBUG uploader snapshotted directory {“path”:“.”,“dur”:“9.068282834s”,“size”:133713551971,“files”:577889,“dirs”:22960,“errors”:1}
BTW, I love that the message uses proper SI-units (µs instead of us)
. Well done.
So this looks like a locking/exclusive read problem. Is there a way to use the APFS snapshotting feature of MacOS? That would resolve these kinds of problems.
Looking through the log I see just these two problems repeating every hour. I will “fix” the second by adding an Exclude rule. How can I troubleshoot the first one (the connection problem)? Did it succeed or not?
P.S. Needless to say, the user feedback can be a bit improved upon. It would be great if the log could end with a conclusion: backup successful / only partially backed up due to errors / etc.