has anyone first hand experience with using live #zfs snapshots of #postgres? what i read seems to confirm that postgres should just start right up like it was recovering from a crash. what i worry about a bit is a quite large and high traffic database thats currently being handled by pgbackrest (which even for diff backups is quite slow).
currently considering something like ZnapZend :think_bread:
background being two dead ssd raids after reboot in a row.
What is the fastest way to walk a filesystem on #Linux?
I am looking to replace locate by enumerating my home directory into a #postgres databases. In a PoC, searching was extremely fast, but initial ingestion and updates are costly operations.
Maybe there's a way to speed things up? Maybe I should read the block device directly with a userspace implementation? Sounds tricky, potentially dangerous and reduces compatibility. Mhmm 🤔
FerretDB enables you to run MongoDB workloads on #Postgres and #SQLite.
Since Neon maintains 100% compatibility with any application that uses the official Postgres release, that makes #Neon an excellent option to consider as the database backend for FerretDB.