Allan Engelhardt <allane@cybaea.com> writes:
> Is there a known problem with HASH type index in PostgreSQL 7.1.2
> 4PGDG on Red Hat Linux 7.1 (2.4.2 kernel)?
It's got a number of known shortcomings, but not anything like the one
you describe. I couldn't reproduce it. I ran your script, then did
VACUUM VERBOSE and VACUUM VERBOSE ANALYZE:
NOTICE: --Relation clients--
NOTICE: Pages 23256: Changed 0, reaped 0, Empty 0, New 0; Tup 1000000: Vac 0, Keep/VTL 0/0, Crash 0, UnUsed 0, MinLen
167,MaxLen 183; Re-using: Free/Avail. Space 2345632/1600; EndEmpty/Avail. Pages 0/1. CPU 4.37s/0.95u sec elapsed 55.65
sec.
NOTICE: Index clients_idx: Pages 6614; Tuples 1000000. CPU 1.18s/35.89u sec elapsed 59.93 sec.
NOTICE: Rel clients: Pages: 23256 --> 23256; Tuple(s) moved: 0. CPU 0.00s/0.00u sec elapsed 0.00 sec.
VACUUM
NOTICE: --Relation clients--
NOTICE: Pages 23256: Changed 0, reaped 0, Empty 0, New 0; Tup 1000000: Vac 0, Keep/VTL 0/0, Crash 0, UnUsed 0, MinLen
167,MaxLen 183; Re-using: Free/Avail. Space 2345632/1600; EndEmpty/Avail. Pages 0/1. CPU 4.41s/0.96u sec elapsed 25.78
sec.
NOTICE: Index clients_idx: Pages 6614; Tuples 1000000. CPU 1.16s/35.85u sec elapsed 62.99 sec.
NOTICE: Rel clients: Pages: 23256 --> 23256; Tuple(s) moved: 0. CPU 0.00s/0.00u sec elapsed 0.00 sec.
NOTICE: Analyzing clients
VACUUM
Looks fine here.
regards, tom lane