On 07/03/2011 01:00 PM, Geoffrey Myers wrote:
> We have a process that we successfully ran on virtually identical
> databases. The process completed fine on a machine with 8 gig of
> memory. The process fails when run on another machine that has 16 gig
> of memory with the following error:
>
> out of memory for query result
>
> How is this possible?
Look at the diff on the postgresql.conf from the two machines.
>
> The databases are almost identical. By that I mean, the database that
> the process completed properly is a dump of the database from the
> machine where it failed. There is about a week's more data in the
> database where the process failed. The whole database only takes up
> about 13 gig of disk space.
>
> Any clues would be appreciated.
>