I found a shell command to look at TM logs, and it came up with this:
The next time you're able to reproduce the issue, my suggestion would be to trigger a sysdiagnose, then file a bug, and finally post the bug number back here. I have a few comments on the logging you sent, but it's possible the full log will show something more definitive.
Looking at the log messages themselves:
2026-09-13 14:03:33 fs_snapshot_list failed: Operation not supported
This means "fs_snapshot_list" returned ENOTSUP (errno "not supported"), which basically only happens... if the target volume doesn't support snapshots. I suspect this is actually just noise caused by activity targeting a different volume unrelated to the current operation.
2026-09-13 14:02:51 Copy failed because of a problem on destination, error: -54
Error "-54" is Carbon's "permErr". TimeMachine is using the same underlying copy engine that the Finder uses, which uses Carbon error codes because that was used when it was originally written and there’s never been any reason to change its error code system, even though the engine itself has changed considerably over time.
2026-09-13 14:02:51 To destination: '/Volumes/TB RAID/../Music Sources/Tomkins', stats: -
That's the destination directory itself, NOT the specific file. In other words, at least as from what the error shows, the problem here isn't that it can't access the source file, it's that it can't create the new file at the destination. WHY it can't... is a question I can't answer from the information at hand.
2026-09-13 14:02:51 Got non-OSStatus error Error Domain=com.apple.backupd.ErrorDomain Code=11 "(null)" UserInfo={MessageParameters=(
2026-09-13 14:02:51 Failed copy from volume "Data"
These are both secondary failures, caused by the original failure above.
__
Kevin Elliott
DTS Engineer, CoreOS/Hardware