"Carrington, Matthew (Produban)" <Matthew.Carrington@Produban.co.uk> writes:
> As its AIX I don't have top but using topas and comparing it to other processes running a successful pg_dumpall
doesn'tget very large at all.
Hmm. Best guess at this point is that there's some specific DDL in your
database that confuses some recent change in pg_dump. Can you try to
narrow down what it is? Try taking a pg_dump -s (schema only) from
the cahoot_monitoring database, load parts of that into a scratch
database, see if 9.2 pg_dump fails on that. Alternatively, if there's
nothing terribly sensitive in your DDL, maybe you could send me the
pg_dump -s output off-list?
regards, tom lane