On 30 January 2017 at 16:34, Peter Eisentraut <peter.eisentr...@2ndquadrant.com> wrote: > On 1/30/17 9:04 AM, Simon Riggs wrote: >> all I want in this release is >> super-ownership. > > What exactly is super-ownership, and what problems does it solve?
The problem is that there is no easy way for a DBA to have privs on multiple sets of objects, so there is a request for superuser in many cases. Superuser is too strong for most situations, so we are stuck. We need some middle ground where a single user can manage many "normal application objects" (tables, views, sequences, matviews, functions, indexes, triggers) without problem, while not compromising other areas that require higher security. Probably more than 50% of PostgreSQL installs now use services that block superuser accounts, so the majority of PostgreSQL users are affected by these problems. The permissions desirable for logical replication are a good example of this, but not in any sense the only issue. My hope is that we release v10 with a permissions model that allows logical replication to be realistically usable when superuser is not available. This is not a new requirement, but the privilege aspect of the logical replication has been pushed back. While thinking about other problems of access control I've rethought this so I now see the wider problem and would like to solve that rather than just focus on the needs of logical replication. -- Simon Riggs http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
