> 1. Do a pg_dump on server1 (it has the space) and then backup the dump file
> 2. Do the same on server2 instead
> 3. Set up a PITR backup on server2
> 4. Do a simple file system backup. This requires shutting down the
> database while the backup is in progress.
They are all valid used solutions to backup postgresql databases. For option 3 i think you mean the hot postgresql backup, witch requires enabling db archiving, before job scripts in order to put database in backup mode and after job script to quit backup mode.
This method is good for not needing aditional dump generation space, to keep the database available for some commands while doing backup and to allow pitr.
There is a PostgreSQL plugin for Bacula available which does PITR backup and recovery. It automatically handles WAL archiving and online datafiles backup. The plugin is available with AGPLv3 license.
It is especially useful for large databases where data import will take a long time or when Point-In-Time-Recovery is required.
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
