Re: Very slow bytea extraction

>Hmm ... what does VACUUM VERBOSE of that table show?
>

Here is the
output of VACUUM VERBOSE

INFO:  vacuuming "public.FILE"
INFO:  scanned
index "FILE_pkey" to remove 54 row versions
DETAIL:  CPU 0.00s/0.00u
sec elapsed 0.01 sec.
INFO:  "FILE": removed 54 row versions in 5 pages
DETAIL:  CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO:  index "FILE_pkey"
now contains 34 row versions in 2 pages
DETAIL:  54 index row versions
were removed.
0 index pages have been deleted, 0 are currently
reusable.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO:  "FILE": found 54
removable, 34 nonremovable row versions in 5 pages
DETAIL:  0 dead row
versions cannot be removed yet.
There were 0 unused item pointers.
5
pages contain useful free space.
0 pages are entirely empty.
CPU 0.00
s/0.00u sec elapsed 0.07 sec.
INFO:  vacuuming "pg_toast.
pg_toast_45181"
INFO:  index "pg_toast_45181_index" now contains 16608
row versions in 51 pages
DETAIL:  0 index row versions were removed.
0
index pages have been deleted, 0 are currently reusable.
CPU 0.00s/0.
00u sec elapsed 0.26 sec.
INFO:  "pg_toast_45181": found 0 removable,
16608 nonremovable row versions in 4
153 pages
DETAIL:  0 dead row
versions cannot be removed yet.
There were 0 unused item pointers.
2
pages contain useful free space.
0 pages are entirely empty.
CPU 0.23
s/0.07u sec elapsed 4.93 sec.
VACUUM

Regards

Massimo