I'm setting up PITR for a client and have a few questions.
I have done some googling for real world archive_command examples and
haven't really found anything. The example in the PGSQL Docs are
qualified by (This is an example, not a recommendation, and may not work
on all platforms.)
I have it set as follows:
archive_command = 'rsync -a %p backup_server:/pgsql_pitr/%f'
Any comments as to whether or not this is a *good* choice?
Also, I'm concerned that this clients website has extended periods of
time where it's very low traffic, which will result in the same WAL file
being used for long periods of time and not getting archived. Does
anyone have a tested script available for grabbing the most recent WAL
file? I can write one myself, but it seems this is information that
should be posted somewhere.
Thanks,
