Thread: Need to Reload a Dump File - Need to Delete the PGDATA directory Again?

Hi All,

I'm in the processing of migrating to Postgres 8.3.8.
I did  :
(1) initdb
(2) psql -e template1 -f 21.bak &>/tmp/out21.bak

Now, I've  decided that I want to reload the same file again, do I just delete the PGDATA directory and then
perform the same steps again?
(1) initdb
(2) psql -e template1 -f 21.bak &>/tmp/out21.bak

What I'm asking is that do I always need to delete the PGDATA directory and then initdb to ensure that it will be a
cleanload? 

Thanks
Mary

Wang, Mary Y wrote:
> Hi All,
>
> I'm in the processing of migrating to Postgres 8.3.8.
> I did  :
> (1) initdb
> (2) psql -e template1 -f 21.bak &>/tmp/out21.bak
>
> Now, I've  decided that I want to reload the same file again, do I just delete the PGDATA directory and then
> perform the same steps again?
> (1) initdb
> (2) psql -e template1 -f 21.bak &>/tmp/out21.bak
>
> What I'm asking is that do I always need to delete the PGDATA directory and then initdb to ensure that it will be a
cleanload? 
>

is that dumpfile from pg_dumpall or pg_dump of a single database or what?

if its just a single database, it should suffice to drop database mydb;
create database mydb with owner myowner;   and then run the script again.

if its the full cluster dump, then yeah, deleting the $PGDATA and
redoing initdb is the best way to insure clean initial conditions.