Patrick Welche wrote: <snip>
Is there more to "remote control" than setting GUC variables? Tell me more!
Sure:
+ starting/restarting/stopping the database server software
+ the normal DBA type work - creating/altering/dropping databases, users, functions, languages, permissions (etc)
+ Remote backup and recovery
+ Submitting jobs to run remotely on the server. i.e. reindexing or vacuuming scripts
Remote Monitoring:
+ Alerts for specified events. i.e. The database server is getting near to capacity in it's filesystem(s), or there have been too many invalid PG authorisation attempts, or there are connections getting rejected because the max_connections parameter isn't high enough
Groups
+ Defining arbitrary groups of servers for the above to speed things up when working with many servers
Roles
+ Having multiple administrators with different permissions (role based is generally good), all communicating through the centralised info repository so things don't get out of sync
(possibly)
+ loading additional PG packages. i.e. rolling out "oid2name" or "pgbench" (or other PG utils) to servers. Could be viewed as something that should be done with the OS packaging mechanism(s) instead.
Any of the PG GUI's (I generally use pgAdmin) could likely be extended to do all of this in a nice, user friendly way.
As an aside, SNMP is important in enterprise settings as it allows PG to be plugged into the monitoring capabilities of enterprise management frameworks. i.e. Concord's eHealth, and probably Tivoli, OpenView, etc
Hope that's useful.
Regards and best wishes,
Justin Clift
Cheers,
Patrick
---------------------------(end of broadcast)--------------------------- TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
