Tom Lane wrote:
Cott Lang [EMAIL PROTECTED] writes:
Fiddling with the above values, only setting sort_mem absurdly large
easily causes NAN.
Ah. I see an overflow case for sort_mem exceeding 1Gb; that's probably
what you tickled.
I've fixed this in HEAD, but it doesn't seem worth back-patching.
Cott Lang [EMAIL PROTECTED] writes:
- Sort (cost=nan..nan rows=2023865 width=1257)
What PG version is this? My recollection is we fixed such a thing quite
some time ago ...
regards, tom lane
---(end of broadcast)---
Oops, sorry - guess I left that out - 7.4.5. :)
On Fri, 2004-10-22 at 12:28, Tom Lane wrote:
Cott Lang [EMAIL PROTECTED] writes:
- Sort (cost=nan..nan rows=2023865 width=1257)
What PG version is this? My recollection is we fixed such a thing quite
some time ago ...
Cott Lang [EMAIL PROTECTED] writes:
Oops, sorry - guess I left that out - 7.4.5. :)
Hmm ... I can't duplicate any misbehavior here. Are you using
nondefault values for any planner parameters? (particularly sort_mem,
random_page_cost, effective_cache_size)
regards, tom
shared_buffers = 16384
sort_mem = 8192
random_page_cost = 2
effective_cache_size = 3932160
On Fri, 2004-10-22 at 13:32, Tom Lane wrote:
Cott Lang [EMAIL PROTECTED] writes:
Oops, sorry - guess I left that out - 7.4.5. :)
Hmm ... I can't duplicate any misbehavior here. Are you using
Cott Lang [EMAIL PROTECTED] writes:
sort_mem = 8192
random_page_cost = 2
effective_cache_size = 3932160
effective_cache_size 30Gb ? Seems a tad high ;-)
However, I set up a dummy test case on 7.4.5 and don't see any overflow.
regression=# create table z1(f1 char(1253));
CREATE TABLE
On Fri, 2004-10-22 at 14:19, Tom Lane wrote:
Cott Lang [EMAIL PROTECTED] writes:
sort_mem = 8192
random_page_cost = 2
effective_cache_size = 3932160
effective_cache_size 30Gb ? Seems a tad high ;-)
It's a 32GB machine with nothing else running on it except PG, buffers
hover around 31GB
Cott Lang [EMAIL PROTECTED] writes:
Fiddling with the above values, only setting sort_mem absurdly large
easily causes NAN.
Ah. I see an overflow case for sort_mem exceeding 1Gb; that's probably
what you tickled.
I've fixed this in HEAD, but it doesn't seem worth back-patching.
If you care,