What about upping ulimt(3) via ulimit(1) in builtin(1) -- assuming you're
running Bash.
~BAS
On Mon, 9 Jan 2006, Ari Kahn wrote:
> I'm doing a query that really should be too taxing. But when I execute it I
> get the following error(s):
>
> psql(18967) malloc: *** vm_allocate(size=8421376) failed (error code=3)
> psql(18967) malloc: *** error: can't allocate region
> psql(18967) malloc: *** set a breakpoint in szone_error to debug
>
> EXPLAIN ANALYZE SELECT a1.qname, a1.symbol, a1.num, a1.ge, a1.start, a1.stop,
> a1.cr, a1.str, a1.ex
> FROM singlehits a1, singlehit_ge a2
> WHERE a2.cnt>1 AND a2.symbol=a2.symbol;
> QUERY PLAN
>
------------------------------------------------------------------------------------------------------------------------------
> Nested Loop (cost=89.36..6086.42 rows=273312 width=88) (actual
> time=0.113..24456.508 rows=54952794 loops=1)
> -> Seq Scan on singlehits a1 (cost=0.00..530.82 rows=17082 width=88)
> (actual time=0.043..71.127 rows=17082 loops=1)
> -> Materialize (cost=89.36..89.52 rows=16 width=0) (actual
> time=0.000..0.418 rows=3217 loops=17082)
> -> Seq Scan on singlehit_ge a2 (cost=0.00..89.34 rows=16 width=0)
> (actual time=0.049..5.167 rows=3217 loops=1)
> Filter: ((cnt > 1) AND ((symbol)::text = (symbol)::text))
> Total runtime: 30024.664 ms
> (6 rows)
>
> Given this post http://xy1.org/pgsql-general@postgresql.org/msg01154.html
> I tried both VACUUM FULL and ANALYZE on the DB to no avail.
> Thanks,
> Ari
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: explain analyze is your friend
l8*
-lava
x.25 - minix - bitnet - plan9 - 110 bps - ASR 33 - base8