[Geoserver-users] Geoserver: displaying 956, 157 points causes Apache Tomcat to be very unstable

Hi,

Listed below are the times for displaying 956,157 points within Geoserver:

  • Geoserver → PostGIS/PostGres: Initially 88 seconds then each time after (after clearing the client side cache) 40 seconds

  • Geoserver → Shapefile: Initially 15 seconds then each time after (after clearing the client side cache) 8 seconds

When using Geoserver version 1.6.4 to display 956,157 points it causes Apache Tomcat version 6.0 to be very unstable. Apache Tomcat would stop after 6-10 Geoserver queries against this 956,157 point dataset both when pointing to PostGIS/PostGres and Shapefile.

What could be done to make these queries more stable like configuration to Geoserver, Apache Tomcat or the server that both Geoserver and Apache Tomcat reside on.

Thanks,

John


John J. Mitchell

What hardware is being used, and in particular, how much memory is allocated for Tomcat's JVM?

When you say "Apache Tomcat would stop" what exactly do you mean? A segfault? A low-level Java exception? Freezing behavior?

---
A. Soroka / Digital Scholarship Services R & D
the University of Virginia Library

On Dec 7, 2008, at 9:21 AM, John Mitchell wrote:

Hi,

Listed below are the times for displaying 956,157 points within Geoserver:

* Geoserver -> PostGIS/PostGres: Initially 88 seconds then each time after (after clearing the client side cache) 40 seconds

* Geoserver -> Shapefile: Initially 15 seconds then each time after (after clearing the client side cache) 8 seconds

When using Geoserver version 1.6.4 to display 956,157 points it causes Apache Tomcat version 6.0 to be very unstable. Apache Tomcat would stop after 6-10 Geoserver queries against this 956,157 point dataset both when pointing to PostGIS/PostGres and Shapefile.

What could be done to make these queries more stable like configuration to Geoserver, Apache Tomcat or the server that both Geoserver and Apache Tomcat reside on.

Thanks,

John

--
John J. Mitchell
------------------------------------------------------------------------------
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you. Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

ajs6f wrote:

What hardware is being used, and in particular, how much memory is allocated for Tomcat's JVM?

When you say "Apache Tomcat would stop" what exactly do you mean? A segfault? A low-level Java exception? Freezing behavior?

---

Well ... I Think I face same problem here.
My PostGis have huge records of PoI and Line ... and small number of Polygon.
My CPU is "clasic" P4 and RAM of 2 GB.

I have "Group layer" of PoI, Line, And Polygon.
Single request agains that group will take almost 90% of CPU and more than 90% of RAM, also more than 10 minutes time to render.

Note : Version is 1.7.0a

Sincerely
-bino-

Bino Oetomo ha scritto:

ajs6f wrote:

What hardware is being used, and in particular, how much memory is allocated for Tomcat's JVM?

When you say "Apache Tomcat would stop" what exactly do you mean? A segfault? A low-level Java exception? Freezing behavior?

---

Well ... I Think I face same problem here.
My PostGis have huge records of PoI and Line ... and small number of Polygon.
My CPU is "clasic" P4 and RAM of 2 GB.

I have "Group layer" of PoI, Line, And Polygon.
Single request agains that group will take almost 90% of CPU and more than 90% of RAM, also more than 10 minutes time to render.

I never tried this exercise (rendering a ton of points) so I have
no specific suggestions. There is a person on gt2-users that
created a custom renderer exactly for this case:
http://conference.osgeo.org/index.php/foss4g/2008/paper/view/340/196

Unfortunately it does not integrate with the GeoTools main renderer at
all, we'd need to redesign the whole thing to allow for pugglable,
context specific renderers.

Cheers
Andrea

--
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.