Цитат от Robert Haas <robertmhaas@gmail.com>:
> 2009/9/14 <tv@fuzzy.cz>:
>> It seems there's something very wrong - the plans are "equal" but in the
>> first case the results (actual time) are multiplied by 100. Eithere there
>> is some sort of cache (so the second execution is much faster), or the
>> system was busy during the first execution, or there is something wrong
>> with the hardware.
>
> I think you should run this query more than twice. If it's slow the
> first time and fast every time for many executions after that, then
> it's probably just the data getting loaded into the OS cache (or
> shared buffers). If it's bouncing back and forth between fast and
> slow, you might want to check whether your machine is swapping.
I did it many times. Alter the first atempt it works fast, but after a
couple of minutes ( I think after changing the data in cache) the
query is working also very slow.
I do not see any swap on OS.
>
> It might also be helpful to post all the uncommented settings from
> your postgresql.conf file.
postgresql.conf :
max_connections = 2000
shared_buffers = 1800MB
temp_buffers = 80MB
work_mem = 120MB
maintenance_work_mem = 100MB
max_fsm_pages = 404800
max_fsm_relations = 5000
max_files_per_process = 2000
wal_buffers = 64MB
checkpoint_segments = 30
effective_cache_size = 5000MB
default_statistics_target = 800
All the rest are default parameters.
Ivan.
>
> ...Robert
>
> --
> Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-performance
>
>
-------------------------------------
Powered by Mail.BG - http://mail.bg