abdujaparov schrieb am 24.09.2015 um 16:39:
> This is my problem.
>
> We must load a lot of data in a postgresql database.
>
> We tried to load 380000 rows in a not partitioned table on postgresql. To
> load this data we spent about 1,5 hours. The table has a serial field and
> the sequence cache is 10000.
>
> After that we tried to load the same rows (380000) on a partitioned table.
> To load data we spent about 2,5 hours. The table is the same (serial filed
> and sequence cache is 10000). Following the postgresql manual to implement
> partitioning I created a father table, child tables, function to insert data
> in the correct child table and the trigger that calls the function.
>
> In each cases I must monitor sessions on postgresql and I do not know how to
> monitor.
>
I just did a bulk load of well over 500000 rows using "copy from stdin" in roughly 30 seconds.
(The source file is > 600MB)
So there must be something off with the way you load the data.
How exactly are you loading those rows?
How many indexes do you have on the table?
Did you turn off autocommit?
You also usually don't need to set the sequence cache that high.
Thomas