On Aug 3, 2013, at 24:04, BladeOfLight16 <bladeoflight16@gmail.com> wrote:
> My big concern as I've been reading this thread is whether users are cluster specific or installation specific. If
they'recluster specific, he'll need to know credentials for his original cluster anyway to get the data, unless he can
dosome kind of password reset.
They are cluster specific, as the roles are stored in the database. If you switch between different data directories,
thatmeans you're switching the available roles as well. And their details, such as passwords. You're also switching
betweenconfigurations, such as pg_hba.conf.
Where it gets confusing a bit here is that there's usually also an OS postgres user, but that user is used to run the
postgresserver/service with limited credentials and not to log into the database. On Windows, apparently postgres is
rununder a standard network service account instead, which serves the same purpose.
Now, if you connect to the database without specifying a role-name to connect as, the standard tools (psql, pg_dump,
etc.Don't know about pgAdmin) take your user account name and try to use that for the database login role, making it
looklike there's a relation between OS users and database users, but that's not actually the case (although there's an
authenticationoption in pg_hba.conf to require such a relation).
At least, this is how I think it works. If I'm wrong someone will no doubt correct me ;)
Alban Hertroys
--
If you can't see the forest for the trees,
cut the trees and you'll find there is no forest.
