[postgis-users] help needed with GEOS & other errors on postgis query

[strk at refractions.net]

PostgreSQL has managed memory creating memory contexts with a defined
life time. We usually do our work in a query memory context. When
the query is complete all allocated memory is released.
If you make multiple queries memory is released earlier then if you
make a single big query.

I've improved early release of memory in upcoming 1.0-RC3 release.

--strk;



On Tue, Feb 22, 2005 at 11:10:55AM -0800, Brent Wood wrote:
> 
> --- strk at refractions.net wrote:
> 
> > There might be a memory leak in the length function.
> > What PostGIS version are you using ?
> > --strk;
> 
> 0.8 from memory.
> 
> To narrow down where the error was happening I tried:
> 
> select distinct set_date from sets; > date.txt
> 
> while read date ; do
> 
> psql ... (same sql but includes where set_date = $date )  >> output
> 
> done < date_txt
> 
> 
> This iterated through a few hundred dates & completed with all the length &
> overlay apparently working fine.
> 
> So you may be right, and reducing the numer of records (by about 1:300) per
> iteration for PostGIS/Postgres cut down the memory issue.
> 
> I've got the answer I needed without any error msgs, but the single query still
> fails, & I can't see why. Both the GEOS & length operations are the same in
> both cases, the only difference I can see is the number of records involved in
> each case.
> 
> 
> Leave it for now or file a bug report? (I'm trying to get our IT staff to
> upgrade the versions of everything, but it's an uphill battle!!)
> 
> 
> Thanks,
> 
>   Brent Wood
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users