On 2013-09-02 19:10:57 -0700, Jeff Davis wrote:
> There is a significant amount of code supporting the changing of
> full_page_writes while the server is running, including an
> XLOG_FPW_CHANGE wal record. But I don't see the use-case; surely almost
> nobody changes this on a running server, because either your filesystem
> guarantees atomic page writes for you or not.
>
> I'd like to submit a patch to just make it into a PGC_POSTMASTER and
> remove the code to support changing it.
I've wondered about that before. But always decided that we wouldn't
gain all that much by making it PGC_POSTMASTER since the effective value
of full_page_writes still needs to be changeable at runtime since we
need to force-enable full page writes during a pg_start/stop_backup().
Greetings,
Andres Freund
-- Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services