[postgis-users] Configuration and performance of PostGIS

[Marco Boeringa]

You are comparing apples and pears.

 From a quick scan at the links you posted, the LineCombiner does a 
fundamentally different thing than the "network-walking" example, it 
simply "combines" any line connected to any other line at its end 
vertex, which is not the same as "network-walking". This is more like a 
PostGIS "ST_Collect()" operation, that also doesn't care about network 
connectivity.

By the way, I never see PostgreSQL take full CPU (2x14 core 
workstation), even with PostgreSQL configured to allow as many CPU cores 
for parallel operations as there are physically. I wish it did, it would 
speed up some operations that really need it. The only way I can get 
full CPU usage is through a custom build Python multi-threading 
framework putting PostgreSQL and PostGIS to work.

Maybe your server is underpowered for the type of work you wish to do.

Marco

Op 22-4-2022 om 01:19 schreef Shaozhong SHI:
> Whenever geospatial functions such as St_intersects or recursive query 
> used, the PostGIS appears to spawn away to many child queries and just 
> obliterate the CPU.  Nothing finishes.
>
> That forced me to try out to do the some tasks on the FME server.
>
> I tried to use this 
> http://blog.cleverelephant.ca/2010/07/network-walking-in-postgis.html 
> <http://blog.cleverelephant.ca/2010/07/network-walking-in-postgis.html> in 
> the PostGIS.
>
> I tried to linecombiner in FME. LineCombiner | FME (safe.com) 
> <https://www.safe.com/transformers/line-combiner/>.
>
> With a large data set, the running of processors were monitored.  It 
> was estimated the PostGIS one would take 16 days to complete.
>
> But, it only took a few minute to do the same thing in FME.
>
> This suggests that something is not right with the PostGIS Server.
>
> Have anyone got experience with configuration and improving perfomance 
> of PostGIS Server?
>
> Regards,
>
> David
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20220422/7f283319/attachment.html>