Achilleus Mantzios wrote:
> What does swapinfo before/after the select show??
Before:
bash-2.05b# swapinfo
Device 1K-blocks Used Avail Capacity Type
/dev/ar0s1b 1048448 296 1048152 0% Interleaved
After:
bash-2.05b# swapinfo
Device 1K-blocks Used Avail Capacity Type
/dev/ar0s1b 1048448 296 1048152 0% Interleaved
>
>
> Also what does vmstat show during the execution of the query?
I run it twice dureing query execution:
1. bash-2.05b# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad4 ad6 in sy cs us sy id
2 0 0 254024 312564 82 0 0 0 49 12 0 0 659 231 46 2 0 98
2. bash-2.05b# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad4 ad6 in sy cs us sy id
2 0 0 500124 60492 82 0 0 0 49 12 0 0 659 231 46 2 0 98
>
>
> Furthermore is the join near a cartesian product??
> (In that case the output relation whould be rather big)
Sorry, but I don't understand what "join near a cartesian product" mean, but
output is only _one_ row since query return only aggregates:
SELECT count(i.*),sum(i.param1),sum(60-i.param1)
>
>
> Im curious since i run freebsd myself.
Our company use FreeBSD for many years (and 4.7 version tested very well).
I think it's not FreeBSD problem, but PostgreSQL. We use dedicated server for
PostgreSQL, so I don't understand why 1Gb of physical RAM exhausted for this simple query...
--
best regards,
Ruslan A Dautkhanov rusland@scn.ru