Hello Freenas Forum.
I have a main machine which is backed up via replication (send/receive) to another machine. They are on the same LAN and in my apartment.
Both machines are on 11.2.
I have configured this as no deletion of stale snapshots, no encryption and no compression.
This has worked quite flawlessly for almost a year.
However... I found out today that alot of data is missing on the backup machine. Several datasets are missing completely. Almost every snapshot is missing, when running "zfs list -t snapshot" on the backup machine (they don't show up in the GUI either).
No data is missing from my machine though.
It seems that the whole replication chain "started over" sometime during the last days.
I did reconfigure the Periodic snapshots on my main machine. I used to recursively snapshot my whole main pool. These snapshots were then replicated to my backup machine (replication task in GUI). My new configuration does not snapshot the whole main pool. I snapshot individual datasets only. I didn't change the replication task though. It was still set to send snapshots of Main pool (which wasn't snapshotted any more).
I assume this has caused the loss of data. My question is Why? Is this expected behavior? Stale snapshots wasn't supposed to be deleted!
Could this have been avoided somehow? I don't use snapshots on my backup machine, I assumed this wasn't necessary as it would be "snapshot of a snapshot", but is this a good idea?
Please help me understand what has happened and how I can avoid this in the future.
Thanks in advance!
I have a main machine which is backed up via replication (send/receive) to another machine. They are on the same LAN and in my apartment.
Both machines are on 11.2.
I have configured this as no deletion of stale snapshots, no encryption and no compression.
This has worked quite flawlessly for almost a year.
However... I found out today that alot of data is missing on the backup machine. Several datasets are missing completely. Almost every snapshot is missing, when running "zfs list -t snapshot" on the backup machine (they don't show up in the GUI either).
No data is missing from my machine though.
It seems that the whole replication chain "started over" sometime during the last days.
I did reconfigure the Periodic snapshots on my main machine. I used to recursively snapshot my whole main pool. These snapshots were then replicated to my backup machine (replication task in GUI). My new configuration does not snapshot the whole main pool. I snapshot individual datasets only. I didn't change the replication task though. It was still set to send snapshots of Main pool (which wasn't snapshotted any more).
I assume this has caused the loss of data. My question is Why? Is this expected behavior? Stale snapshots wasn't supposed to be deleted!
Could this have been avoided somehow? I don't use snapshots on my backup machine, I assumed this wasn't necessary as it would be "snapshot of a snapshot", but is this a good idea?
Please help me understand what has happened and how I can avoid this in the future.
Thanks in advance!
Last edited:
