Time Machine "Unknown Error": Case-Sensitive APFS Destination (and Encrypting It Correctly)¶
Overview¶
Time Machine will refuse a backup destination whose volume is formatted Case-sensitive APFS, but it reports only the generic dialog "Time Machine couldn't complete the backup to '\<disk>' — An unknown error occurred." Nothing in the UI names the filesystem format, so this is easy to misdiagnose (and re-diagnose) for months.
The only fix is to reformat the destination as Case-insensitive APFS (or HFS+), which erases it. If your Mac's startup disk is FileVault-encrypted, you'll then be prompted to encrypt the backup too — and the obvious GUI path to do that silently reformats the volume back to case-sensitive, undoing the fix. This article covers the whole correct sequence, plus the ownership/mount gotcha that blocks re-registering an encrypted destination.
[!warning] The disk gets erased Reformatting the destination destroys whatever is on it. If Time Machine was failing, the existing contents were never a valid backup chain anyway — but confirm you have no other data on that drive first.
Symptoms¶
- "Time Machine couldn't complete the backup to '\<disk>' — An unknown error occurred."
- Backups never complete;
tmutil statusmay reachCopyingbriefly but the job never finishes, or fails immediately. - The destination is mounted, writable, and has plenty of free space (not full, not disconnected).
- Orphaned local snapshots pile up on the source disk (
tmutil listlocalsnapshots /) because they can never transfer.
Diagnosis — check the volume's case sensitivity first¶
This is the fast disqualifier. Run it before anything else:
A healthy destination reads File System Personality: APFS (case-insensitive) or an HFS+ personality. If you see Case-sensitive APFS, that is the root cause — no amount of snapshot cleanup or "remove and re-add destination" will help until it's reformatted.
Fix Part 1 — reformat case-insensitive¶
Identify the physical disk, not the synthesized APFS container. diskutil list shows the container as (synthesized) with a Physical Store pointer — erase the physical device so you get a clean container:
[!note] Device numbers move After every force-unmount or erase, macOS re-synthesizes the APFS container and the
diskNnumbers change. Re-rundiskutil listbetween steps rather than trusting a number you saw earlier.
Remove the old destination, force-unmount (this evicts Spotlight, which otherwise blocks the erase), then erase the physical disk as plain APFS (which is case-insensitive):
sudo tmutil removedestination <old-destination-id> # id from: tmutil destinationinfo
sudo diskutil unmountDisk force /dev/diskN # physical disk
diskutil eraseDisk APFS "TimeMachine" /dev/diskN # "APFS" = case-INsensitive
[!tip] Two harmless-looking errors -
eraseDiskfailing with-69877 … in use by mds_storesmeans Spotlight has the volume open — the forcedunmountDiskabove prevents it. (Alternatively:sudo mdutil -i off "/Volumes/<VolumeName>".) --69842 "Couldn't mount disk … after erase"is usually cosmetic on external USB drives — the volume mounts anyway. Verify withdiskutil infobefore assuming failure.
If you don't need encryption, register the destination and start:
Fix Part 2 — encrypt the destination (source is FileVault-on)¶
If your startup disk is FileVault-encrypted, macOS warns you're "backing up encrypted data to a non-encrypted drive." A portable backup drive absolutely should be encrypted — but do it in the right order.
[!danger] Do NOT use System Settings → Time Machine → "Encrypt Backups" to encrypt the disk That flow erases and recreates the volume as Case-sensitive APFS — straight back to the original root cause (you end up encrypted and broken). The GUI gives you no control over case sensitivity.
Instead, format case-insensitive first (Part 1), then encrypt in place via CLI on the empty volume (near-instant):
diskutil apfs encryptVolume "/Volumes/TimeMachine" -user disk
# prompts for a passphrase — set a strong one and store it in your password manager
The 49229 gotcha after CLI encryption¶
After encryptVolume, the volume typically ends up unmounted with Owners: Disabled. Time Machine requires ownership enabled and a valid mount path, so tmutil setdestination fails with:
That is not a bad-disk error — it means the destination was rejected. Unlock/mount the encrypted volume, enable ownership, then register it:
diskutil apfs unlockVolume diskNsM # unlock + mount (prompts passphrase)
sudo diskutil enableOwnership "/Volumes/TimeMachine" # TM requires ownership enabled
sudo tmutil setdestination -a "/Volumes/TimeMachine"
tmutil startbackup --auto
Verify¶
diskutil info "TimeMachine" | grep -iE "Personality|FileVault|Owners|Mounted"
tmutil destinationinfo
tmutil status | grep -iE "BackupPhase|Running"
You want: Personality: APFS (case-insensitive), FileVault: Yes, Owners: Enabled, a single registered destination, and Running = 1. A first full backup reports BackupPhase = Starting (with Percent = -1) while it enumerates files, then flips to Copying.
Notes¶
- Escrow the passphrase yourself. macOS keeps a copy in the system keychain so the drive auto-unlocks for unattended backups, but keep your own copy in a password manager — it's the only way to read the backup if the keychain is ever lost.
- Across a reboot an encrypted external may need one manual unlock before backups resume; check "Remember password in my keychain" when prompted to make it automatic thereafter.
- This is distinct from the orphaned
.previousfolder failure, which produces the same generic dialog but is fixed with First Aid (fsck_apfs) rather than a reformat. Ifdiskutil infoshows the destination is already case-insensitive, check that article next.