"Ryan D. Enos" <renos@ucla.edu> writes:
> My problem is that my database is creating large duplicate
> files, i.e.: 17398.1, 17398.2, 17398.3, etc. Each is about 1g in size.
These are not "duplicates", they are sections of a very large table
(or possibly a very large index).
I gather from your followup that the files are now gone, meaning that
someone dropped the table. If they were still there you could determine
which table they belonged to by looking at pg_class.relfilenode --- see
http://www.postgresql.org/docs/8.2/static/storage.html
for information about the physical layout of a PG database.
My best guess about what happened is that a client did something like
CREATE TEMP TABLE foo AS SELECT ...
with an erroneous query that generated a huge table --- maybe it had
a missing join condition or something like that.
regards, tom lane