Nik wrote:
> The reason I am running pg_dumpall is that I would like to create an
> exact copy of my database on another server. I cannot use create
> database with template command since there are connections to the
> original database. The dumpall method seems somewhat cumbersome since
> it generates only one file that is huge.
>
> Is there another way I could do this that would be faster, more
> reliable, and elegant? Could I start archiving the logs for the
> original database and then do the pitr on another machine using that
> backup and logs?
>
On the second machine
pg_dumpall -h first_machine | psql
Of course that is assuming the second machine is a unix clone. I don't
know if windows pipe works like unix or if it still works like the old
dos pipe.
