Steve King <steve.king@ecmsys.co.uk> writes:
> I've now got a copy of pg_filedump and compiled it, can you tell me the
> command line parameters to pass it (and the file that I must process) so I
> can give you exactly what you require.
I'd recommend
pg_filedump -f -i -R <blocknum> <filename>
where <blocknum> is whatever page you need to look at (the high part of
the TIDs of the bad tuples), and <filename> is going to be of the form
$PGDATA/base/<dbnum>/<filenum>
You get the DB number from
select oid from pg_database where datname = 'yourdbname';
and the file number from
select relfilenode from pg_class where relname = 'yourtablename';
regards, tom lane
