On Dec27, 2010, at 12:15 , Magnus Hagander wrote:
> Actually, having implemented that and tested it, I realize that's a
> pretty bad idea. For one thing, it broke my own pg_streamrecv program,
> since it requires the ability to connect to the master and select a
> pg_current_xlog_location().
I'm starting to think what we really want here is a kind of read-only
superuser. WITH REPLICATION already essentially gives you read-only
access to the whole database. Thus, allowing WITH REPLICATION roles
read-only access to everything on the SQL level also doesn't really
extend their abilities, it merely makes getting some information faster
and more convenient. It'd also make WITH REPLICATION the perfect fit
for pg_dump-style backups, if you're uneasy about using a superuser
for that.
best regards,
Florian Pflug