Snapshot replication task not following custom retention policy

Hi all,

I have two TrueNAS scale servers, a prod and a backup. Prod has around 20 datasets and I am doing snapshot tasks with replication either on a per hour basis or per day basis (depending on the dataset).

I only want to keep one week worth of snapshots on the prod system as it has less storage than the backup system and then I want to keep most datasets for 1 or 2 months on the backup system but some longer (maybe 6 months or more). I am currently trying to do this using the “Snapshot Retention Policy” on the replication task but this setting does not seem to be respected by the destination system as snapshots are being deleted after 1 week (the lifetime on the prod system) and not the longer lifetime set by the snapshot retention policy. At one point years ago this did work but I am not sure when it stopped working.

I am going to attach some example screenshots of the config that may help but I am trying to figure out why this is not working the way I thought it would and or how to fix it.

Here is an example snapshot task,

Here is an example of a replication task,


Both systems are running Dragonfish-24.04.2 and this issue has existed across multiple versions.

Looking for any suggestions/advice, very possible I have a config issue that I just do not see but I have checked every bit of documentation I can find online about the snapshot retention policy and nothing has shown I am using it incorrectly.

Thanks!

Hello, i have the same problem. Anyone from TrueNAS Team need logs?

I can provide logs if needed as well, just let me know what the team needs.

Really hoping to get this fixed.

Scroll to the top of any forum screen and use the Report a Bug link to report the issue via Jira. After you submit the report and auto comment will provide a link to a private upload service where you can provide a debug file from your system.

Is it PUSH or PULL replication?
If it is PUSH, switch to PULL and see whether having the destination system manage the snapshots achieves what you want.

Ticket NAS-131427 was opened for this issue and debugs from both prod and backup system have been provided.