Not true, you can run the dump from any server that can connect to
the database server (based on network connectivity and the connection
permissions in your pg_hba.conf file). Just be sure to use the -h
(host) -p (port) and -U (username) flags when execute pg_dump.
If you use the -F c (pg_dump -F -c -f your_dmp.file your_db) option
you'll get a compressed dump file which can be as little as 10% (or
less) the size of your database.
If you can stop the database then you can just do a manual copy
(using cp, rsync, or whatever) of the entire pg data directory but
that will require the same amount of space as the original database.
If you're goal is to make a backup, pg_dump is the way to go. Make
note, though, pg_dump will only dump one database so if you have more
than one database in your postgres "cluster" then you'll need to use
pg_dumpall in which case you don't have the compression (-F c)
option. There are other caveats between the two as well. Just be
sure to read the documentation. Also, for other postgres questions,
you should join one of the postgres mailing lists.
Erik Jones
Software Developer | Emma®
(e-mail address removed)
800.595.4401 or 615.292.5888
615.292.0777 (fax)
Emma helps organizations everywhere communicate & market in style.
Visit us online at
http://www.myemma.com