On 10/22/22 17:06, Adrian Klaver wrote:
On 10/22/22 14:45, Ron wrote:
On 10/22/22 16:29, Adrian Klaver wrote:
To pseudo for me.
What file exactly is:
pg_restore --jobs=X --no-owner $NEWDB
restoring?
And how was that file created?
Knowing this might help get at why the more straight forward method does
not work.
This is what I ran to restore the database:
export PGHOST=${RDSENV}.xxxxxxxxxxxx.us-east-1.rds.amazonaws.com
cd /migrate/TASK001793786/2022-10-19b
NEWDB=sides
pg_restore -v --create --clean --no-owner --jobs=`nproc` -Fd
--dbname=template1 $NEWDB
psql $NEWDB -f all_OWNER.sql
psql $NEWDB -f all_GRANT.sql
The name of the database is "sides", and there's a directorynamed "sides"
under /migrate/TASK001793786/2022-10-19b.
Aah, I forgot about the -Fd, now it makes more sense.
To get past the --jobs induced error required the addition of --no-owner
and then adding owners and grants after the main restore.
What was the pg_dump command that produced
/migrate/TASK001793786/2022-10-19b/sides ?
PGHOST=mumble.xxxxxxxxxxxx.us-east-1.rds.amazonaws.com
PGUSER=postgres
cd /migrate/TASK001793786/`date +%F`
DB=sides
pg_dump -d $DB -j4 -Z0 -v -Fd --file=$DB 2> ${DB}_pgdump.log
--
Angular momentum makes the world go 'round.
