Fernando Schapachnik wrote:

>Consider the scripting way:
>
>cmd="SELECT schemaname || '.' || tablename from pg_tables';"
>list=`$psql -c "$cmd" `
>
>i=1
>for k in $list
>do
>        echo "set add table (set id=1, origin=1, id=$i, fully qualified name = 
>'$k');" >> $arch
>        i=`expr $i + 1`
>done
>
>Or something alike.
>
>  
>
That will, by all means, be helpful, although it's more likely to be
useful to do something like...

for db in `psql -t -l | cut -d "|" -f 1`; do
  l=0
  for tbl in `psql -q -d $db -t -c "select schemaname || '.' ||
tablename from pg_tables;"`; do
    l=`expr $l + 1`
    echo "set add table (id=$l, set=1, origin=1, fully qualified
name='$tbl');" >> $db.add.slonik
   done
done

Note that if there are 60 databases, this creates 60 files.  And note
that the above includes a bunch of schemas that probably should be left
out...

It's not bad as a "proof of concept" that it's plausible.

But you then have to configure 60 clusters, subscribe them, and manage
them, which will be not too much fun...