Bug Reports, Operation Guides, and Experience Sharing
https://forum.terra-master.com/en/
Regarding the “1–20 MB fluctuation,” is this occurring during a local network transfer, or is it a remote transfer via some other method?
Based on the screenshot you provided, the version number is unexpectedly 7.0.1341.
Morning, any update on this?storageninja wrote: ↑22 Oct 2026, 14:45 Hi all,
I'm running into a problem with Time Machine backups to a Terra Master NAS over SMB, and I'd like to check whether others have seen this or whether it's a known issue.
Setup
- NAS: Terra Master F4-SSD, TOS TOS7 1341, btrfs volume
- Client: MacBook M4 Pro, macOS 27.0.1
- Backup target: SMB share, Time Machine enabled on the share, with a size quota of 1500 GB set in TOS
- The Time Machine sparse bundle was originally created with a 16 TB virtual size, the default?
What happens
- Backups ran normally until the backup set reached about 1.5 TB on the NAS.
- After that, every backup fails with BACKUP_FAILED_TARGETVOL_DISK_FULL (56), underlying NSPOSIXErrorDomain Code=28 "No space left on device".
- The failure happens at the point where Time Machine opens the mount point, before any data is written:
MountPointValidity - open failed, error: 28 No space left on device
- Old backups are never thinned. The backup set stays above the quota indefinitely, so the problem persists.
- The same behavior happened earlier when the quota was 750 GB and I raised it.
What I checked
- Free space as reported on the client: df on the SMB mount shows about 3.1 TiB free, which matches what the NAS reports for the volume. So the free-space value that macOS sees is not wrong, and it does not show the quota at all.
- The sparse bundle on the NAS is about 1.4 TiB by du, which is what the quota counts against.
- The btrfs qgroup output shows no max_rfer or max_excl limits. So the quota is not a btrfs qgroup limit. It seems to be enforced by a layer in TOS that rejects writes with ENOSPC.
- The Time Machine sparse bundle reports 16 TB capacity internally, and the APFS container inside has ample free space. So the bundle itself is not full.
My reading
Time Machine decides whether there is room (and whether to thin old backups) based on the free space the destination reports. Because the quota is enforced at write/open time but is not reflected in that free-space value, Time Machine sees roughly 3 TB free. *It never starts thinning* and every backup fails at open with ENOSPC, which Time Machine reports as "disk full". As far as I can tell, the quota prevents the backup from ever reaching the point where thinning would run.
Questions
1. Is there a supported way to set a size limit on a Time Machine share on TOS that Time Machine can see (for example, a free-space value that reflects the quota)?
2. Has anyone got Time Machine to thin old backups automatically on a quota-limited share?
3. Is there a recommended approach for capping Time Machine size on Terra Master: remove the quota and reserve space another way, or manage thinning manually with tmutil?
Happy to run any diagnostics. Thanks.
We created a folder named "Küche" and a file named "Küche", and both opened normally in Finder on our test.
Currently, it is not supported to automatically thin out old backups when a Time Machine backup reaches the shared folder quota.
Well thats very bad news, so the only way is to empty the fileshare when it reaches 100%? Thats not even close to the way Apple designed and implemented the TimeMachine look and feel.
We sincerely apologize, but based on your description alone, we are unable to determine the cause of the issue at this time. Additionally, since you have already rebooted the device, the relevant logs from that incident are no longer available. Before the next update, please enable debug mode. If the issue occurs again, please do not reboot the system — contact us directly so we can perform a remote inspection.