PostgreSQL's fsync() surprise

Posted May 2, 2018 23:56 UTC (Wed) by gerdesj (subscriber, #5446)
Parent article: PostgreSQL's fsync() surprise

Perhaps it has been discussed to death before but why not put DBs on some sort of DB oriented storage instead of say xfs/ext{n}/btrfs/fat16?

PostgreSQL's fsync() surprise

Posted May 3, 2018 0:03 UTC (Thu) by andresfreund (subscriber, #69562) [Link] (2 responses)

Which would be?

PostgreSQL's fsync() surprise

Posted May 3, 2018 6:20 UTC (Thu) by zlynx (guest, #2285) [Link] (1 responses)

Raw, unformatted blocks I would suppose. We could call it postgresfs.

PostgreSQL's fsync() surprise

Posted Nov 12, 2018 4:03 UTC (Mon) by immibis (subscriber, #105511) [Link]

That is almost exactly the abstraction that a file is supposed to provide - except that it's a fixed size (and consequently you can't get ENOSPC because you are handling the space allocation yourself). You can still get EIO. Or EUSERPULLEDTHEDRIVEOUT.

PostgreSQL's fsync() surprise

Posted May 3, 2018 11:26 UTC (Thu) by james (subscriber, #1325) [Link]

That means you can only run them on systems with that sort of storage available -- which means
dnf install package-that-uses-postgresql-as-a-database-engine
doesn't have a chance of Just Working.