I performed a snapshot pruning 2 days ago that resulted in some issues on the snapshots. I incurred some light data loss because, I suspect, some snapshots were created as "clone snapshots" and once the original snapshot was deleted, the other snapshots were recursively deleted and as such, the data is gone. I was lucky that some of my jails were running and so a busy snapshot could not be deleted.
The pool is now in a weird state where when I try to perform a replication with the flags "from scratch" and "almost full dataset", from my source (MARS) onto a brand new pool (BMARS), I receive the following errors (this is an extract):
skipping snapshot MARS/mymedia@francesco-202408021221 because it was created after the destination snapshot (francesco-202407311813)skipping snapshot MARS/iocage/jails/znc_1@francesco-202408021221 because it was created after the destination snapshot (francesco-202407311813)cannot send MARS@francesco-202407311813 recursively: snapshot MARS/iocage/jails/znc_1/root@francesco-202407311813 does not existwarning: cannot send 'MARS@francesco-202407311813': backup failedcannot receive: failed to read from stream")Whatever I did, I wasn't able to replicate the pool anymore, and I definitely would like to backup my data.
I don't think the skipping is the source of the issue, but rather the cannot send MARS@francesco-202407311813 recursively: snapshot MARS/iocage/jails/znc_1/root@francesco-202407311813 does not exist. This missing snapshot seems to be the actual source of the problem.
What are my options to correct the error and be able to perform a full replication again?I'm afraid of deleting snapshots and ending up losing data again.
Keep in mind that:
- I was able to perform a full replication right now by creating a new snapshot with a different naming schema and requested to perform the full replication following the new naming schema (
francesco-fix1-202408021221). In this way only one snapshot was replicated (ok), and all the data with it was copied over. This is better than having no backups - I'm concerned that all the "skips" are about snapshots that happened in the future. I definitely want those, because I want that data
- I don't care about previous snapshots as long as I don't lose the data. I'm pretty sure there are more "clone snapshots" around because TrueNAS creates the jails as "clone jails" by default
The problematic snapshot is MARS/iocage/jails/znc_1/root@francesco-202407311813, the most recent snapshot that I would like to have is francesco-202408021221, but as you can see, it says skipping because it was created after the destination snapshot.
Would creating a new snapshot for MARS/iocage/jails/znc_1/root and name it francesco-202407311813 potentially fix the issue, even if the snapshot is created with the current state of the data?