Hi folks,
I want to refer to a question Rob did back in 2008 at
http://archives.postgresql.org/pgsql-general/2008-07/msg01167.php as we
are currently running into a similar question:
We are using warm standby via PITR using a shared drive between master
and slave node.
Our setup currently is set to archive_timeout = 60s and
checkpoint_timeout = 600s.
We expected that now every minute a WAL-file is written to the share,
but somehow we might misunderstood some part of the documentation as in
periods with low traffic on database the interval between WAL files is
>1min up to ten minutes.
However, the goal was to have a WAL file every minute so disaster
recovering can be done fast with a minimum of lost data.
Question is: What did we miss? Do we need to put checkpoint_timeout also
to 60s and does this makes sense at all?
Cheers,
Frank