Thread
-
Terminology gaps around WAL
Jeff Davis <pgsql@j-davis.com> — 2026-06-11T22:14:24Z
I sometimes struggle to come up with the right term for the following concepts, which come up a lot: * The abstract set of WAL records that all originate from a single initdb and share a system identifier; including all timelines, and including old records that may have been deleted a long time ago. * The "stuff" around that specific set of WAL records, such as running instances, warm standbys, PITR'd instances, inactive instances, base backups, and WAL archives. "Database cluster" doesn't fit either of these. The docs define it as "a collection of databases that is managed by a single instance of a running database server." [1] Do others already have good terms for these concepts? Is it worth trying to invent and document some? Regards, Jeff Davis [1] https://www.postgresql.org/docs/current/creating-cluster.html
-
Re: Terminology gaps around WAL
Laurenz Albe <laurenz.albe@cybertec.at> — 2026-06-12T07:24:22Z
On Thu, 2026-06-11 at 15:14 -0700, Jeff Davis wrote: > I sometimes struggle to come up with the right term for the following > concepts, which come up a lot: > > * The abstract set of WAL records that all originate from a single > initdb and share a system identifier; including all timelines, and > including old records that may have been deleted a long time ago. > > * The "stuff" around that specific set of WAL records, such as running > instances, warm standbys, PITR'd instances, inactive instances, base > backups, and WAL archives. > > > "Database cluster" doesn't fit either of these. The docs define it as > "a collection of databases that is managed by a single instance of a > running database server." [1] > > Do others already have good terms for these concepts? Is it worth > trying to invent and document some? > > [1] https://www.postgresql.org/docs/current/creating-cluster.html "Multiverse" pops into my mind. Yours, Laurenz Albe