I use TN as my Time Machine destination for 2 macbooks.
About 2 weeks ago when Fangtooth came out, I upgraded from Eel.
Today I was looking at my snapshots in TrueNAS.
I see that the Time Machine has not taken a snapshot after April 16th.
Normally it takes an automatic snapshot every couple of hours I see.
I use the default Time Machine settings in the TN share. No additional parameters.
Anyone else run into this?
Side note, I have noticed also that a lot of snapshots that had automatic removal turned on (like keep daily snapshots for 3 months), where not automatically deleted anymore. Though they had Retention: “some date ago”. Not sure if that is my local issue or that the upgrade somehow reset a database file that kept track of these. Will keep a track of it. I just wiped a few thousand snapshots to make them up to date with today’s retention date.
(no I didn’t manually remove Time Machine snapshots)
Is anyone else also having this issue that no snapshots are being taken on time machine backups? Or just the 2 of us?
Time machine backups itself work fine, on both macs we have (M1 and Intel running latest Sequoia).
I use multi-user time machine backups.
I created a second multi-user time machine backup share/dataset, no snapshots neither
I also create a third basic time machine backup share/dataset, no snapshots neither.
Tried both macs and also a different username/password.
In all cases the backup completes and also subsequent incremental backups, but I see 0 snapshots. (only the old ones until April 16th; when I was on Eel)
thanks for sharing this. I don’t have permission to view this for some reason. Can you please share what’s the status of bug report? Is there an ETA for a patch?
The status is unclear to me, says Waiting for response.
Andrew Walker is looking at it (last communication May 2025)
I have provided log files.
So no ETA so far.
Thanks, the profile used to do this and that should be the intentional use case. Creating manual snapshots seems to me interfering with the underlying system. For now it is a good work around.
My understanding is that this “bug “ is not due to any change in TrueNAS software.
The SMB-level heuristics used to decide when a Timemachine backup is completed no longer work reliably. We suspect that the TimeMachine software has changed in some subtle ways.
Without a clear backup completion time a useful snapshot cannot be taken.
We’re not sure its worthwhile to try to fix this. Instead, we recommend a periodic snapshot at a time when the timemachine process is inactive.
If there was a clear signal from TimeMachine on completion of a backup job, we could reconsider.
I thought the resolution to this issue was that multi-user SMB Time Machine shares did not have this problem while dedicated TM shares could?
FWIW, I am running multiple dedicated TM shares on 25.04 w/just the usual issues. Ie had a basic TM archive corrupt and refuse to take further updates. Deleted the TM share contents, redid the TM backup from scratch. That was likely a bad choice, however, I should have tried rolling back to an earlier snapshot first before nuking the archive.
So I wonder if the issue can be resolved by rolling back the TM to an earlier state or rebuilding it from scratch? IIRC, the multi-user TM had some extra bits in it that made it more resistant to snapshots being executed while TM was happily writing to the share. Perhaps @awalkerix can elucidate best practices around TM and TrueNAS ATM?
