Sometime early this morning, our database started having issues, most likely due to a disk corruption issue.
Unfortunately the postgres database was unrecoverable, and we had to restore from a nightly backup, so a few hours of posts and comments will be missing.
Apologies again for the downtime, and for the disruption.
How does lemmy federation work in this case? Conceivably after being restored from backup the lemmy.ml instance could see those few hours of lost history as federated to other lemmy instances and resync it back as the host instance. Obv I’m vastly oversimplifiying things but what happens today?
I believe they’re lost, as recieved apub content doesn’t have the original post / comments, but I’m not totally sure. cc @nutomic@lemmy.ml
It strikes me that there is the potential to use trusted remote servers as a means of recovering the lost data. I mean, nearly every lemmy instance except lemmy.ml will have copies of the missing data, and given the hugely redundant availability of that data (including the ability to compare from multiple sources to establish/verify trust), using that data to rebuild missing content seems like it could be useful functionality.
If I understand, federating contents through ActivityPub only works once. Sounds a good feature to re-download contents again, but may introduce additional work as there should be some way to know if a content is missing then another job to rebuild
That doesnt work because we generally cant trust remote servers. Plus we dont even know where to fetch from, so wed have to run a complete crawl of all known instances which isnt practical.