On Fri, May 18, 2018 at 8:13 PM, Stephen Frost <sfrost@snowman.net> wrote:
> I'm not a big fan of it- what happens when we introduce something else
> which would seem like it'd fall under 'maintain' but provides some
> capability that maybe it wouldn't be good for users who could only run
> the above commands to have? I'm tempted to suggest that, really, we
> might even be thinking about splitting up things further than the above
> proposal- what about VACUUM vs. VACUUM FULL? Or REFRESH MATVIEW vs.
> REFRESH MATVIEW CONCURRENTLY? Mistakes between those routinly cause
> problems due to the heavy lock taken in some cases- as an administrator,
> I'd be a lot more comfortable giving a user or some process the ability
> to run a VACUUM vs. VACUUM FULL.
That is a fair point, but if we want to do things like that then it's
really not a good idea to limit ourselves to a fixed number of bits,
even if it's 2x or 4x more than what we have today.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
