I’m using TrueNAS SCALE 25.10.7 and TrueCloud Backup (Storj).
Unfortunately, my Backup jobs are failing sporadically with this error message (from /var/log/jobs/xxx.log)
warning: running prune without a cache, this may be very slow!
loading indexes...
loading all snapshots...
finding data that is still in use for 14 snapshots
[17:04] 100.00% 14 / 14 snapshots
The difference between the request time and the server's time is too large.
ReadFull(<data/5639f4dddc>)
github.com/restic/restic/internal/backend.ReadAt
/restic/internal/backend/readerat.go:39
github.com/restic/restic/internal/repository.(*Repository).LoadBlob
/restic/internal/repository/repository.go:297
github.com/restic/restic/internal/restic.LoadTree
/restic/internal/restic/tree.go:115
github.com/restic/restic/internal/restic.loadTreeWorker
/restic/internal/restic/tree_stream.go:36
github.com/restic/restic/internal/restic.StreamTrees.func1
/restic/internal/restic/tree_stream.go:176
golang.org/x/sync/errgroup.(*Group).Go.func1
/home/build/go/pkg/mod/golang.org/x/sync@v0.5.0/errgroup/errgroup.go:75
runtime.goexit
/usr/local/go/src/runtime/asm_amd64.s:1650
I don’t think that the problem exists because the time on my TrueNAS is incorrect, because it should be fine:
truenas_admin@nas[~]$ timedatectl
Local time: Thu 2026-10-29 07:24:21 CET
Universal time: Thu 2026-10-29 06:24:21 UTC
RTC time: Thu 2026-10-29 06:24:21
Time zone: Europe/Berlin (CET, +0100)
System clock synchronized: yes
NTP service: active
RTC in local TZ: no
These are the settings for my backup job:
Path: /mnt/tank/data/Andy
Cache Path: -
Credentials: Storj
Bucket: backup
Folder: Andy
Keep Last: 14
Rate Limit: No limit
Transfer Setting: Default
Post Script: -
Pre Script: -
What could be causing this problem?
Clause suggests to run restic without --no-cache, but I guess there is a reason why this parameter is enabled by default for TrueCloud Backups, right?
Oh, that’s very unfortunate. Especially because Storj is the only cloud service provider that is compatible with TrueCloud Backup right now? Cloud Sync does not support retention and therefore is no real backup.
Will TrueNAS open up TrueCloud Backup to other services?
IMHO, retention should be set on the destination, not on TrueNAS. Imagine you getting malwared, in that case the attacker could just change your retention settings.
Just use S3. Backblaze B2 for example is cheap and you can set retention (Object lock) in the webGUI.
I don’t know much about how Backblaze B2 works. I understand that with Object Lock enabled, files will not be deleted for e.g. 30 days. Also when files are being overwritten, an older version of the file will be available. But how can I access these files, for example if all my data has been encrypted by ransomware and some of it has been uploaded already?
I did now setup Backblaze B2 with Lifecycle Rule (30 days), so I think the backup should be safe in case of a Ransomware attack or data corruption. I was also able to download the backup at a specific point in time to my computer.
That’s right, but the API Key for TrueNAS has no rights to delete files, it can just mark them as hidden. That’s what rclone does, when a file at the source is deleted. Hidden files are then deleted by the lifecycle rule after 30 days.
Do you use Object Lock or how do you do it? As far as I understood, the Object Lock timer starts at the creation of the file on S3, not when it’s being deleted. That makes it useless for my case.
You’re right about the timer. Object Lock retention is set per file version and counts from when that version was uploaded, not from when it gets hidden. It isn’t useless for you though. When an encrypted copy of a file syncs up, the previous version gets hidden, and if that previous version is still inside its lock period the lifecycle rule can’t remove it. Where it stops helping is files that were uploaded long ago and never changed. Their lock has already run out, so your 30 day hide-to-delete rule is the only thing keeping the old version around.
So the number that really matters is how fast you’d notice. If bad files sync on day 1 and nobody spots it until day 35, the lifecycle rule has already removed the good versions. I’d set daysFromHidingToDeleting comfortably longer than your realistic detection time, and keep an eye on the job log. A run that suddenly uploads or hides a huge number of files is the early warning.
The key with no delete rights is a good move, just keep the master key off the NAS. And when you do need the bucket, pull a point from before the first bad upload into a separate folder and open a sample of files before you overwrite anything on the pool.