Replication resume token is not updated without data writes

This issue has now been closed. Comments made after this point may not be viewed by the TrueNAS Teams. Please open a new issue if you have found a problem or need to re-engage with the TrueNAS Engineering Teams.

We've merged the patch into upcoming SCALE 24.04 and Core 13.3 releases.

I was able to reproduce the scenario, and this slightly hackish patch fixes it for me: https://github.com/openzfs/zfs/pull/15927. Lets see what the community think about it.

I’m glad you found the issue! I’ve deactivated atime now as I don’t need it. I haven’t touched this setting since years. I may have ran a find command on the filesystem, therefore touching every file. I don’t know, but that would be an explanation. However, setting atime should not break replication. But I suppose that’s what you’re fixing right now. Let me know if you need additional infos or if I should try something else. Thanks for taking the time to look into that.

I see the problem. ZFS updates receive_resume_token only when receiving some data. But the stream in this case includes no data writes at all. Without token updates for third-party observer it looks like replication is stuck, and after restart it restart from the beginning again. I need to look for other good point to update the token.