I have following geoserver status. I have installed JAI but memory used by jai is 0 and image rendering for wms is very slow. enter image description here

Is it issue with memory allocation? I am using Sql server 2008.

  • what sort of tiles are you rendering - if they are vector tiles then I suspect JAI is not really involved in the process.
    – Ian Turton
    Feb 24, 2012 at 11:13
  • @iant I am using layer published from SQL server and using WMS. Does this mean it's vector tiles? And if so then performance will not be altered by this??
    – kinkajou
    Feb 24, 2012 at 11:45

2 Answers 2


From this page:

In particular, installing the native JAI is important for all raster processing, which is used heavily in both WMS and WCS to rescale, cut and reproject rasters. Installing the native JAI is also important for all raster reading and writing, which affects both WMS and WCS. Finally, native JAI is very useful even if there is no raster data involved, as WMS output encoding requires writing PNG/GIF/JPEG images, which are themselves rasters.

So if you are drawing WMS tiles of a vector data set you will see some speed up from using the native JAI but probably not much, if your map is a coverage/raster dataset (satellite imagery or aerial photos) then you will see a speed up by using the native JAI.

If you have slow map generation from a vector source the most likely slow downs are that you don't have a spatial index on yyour datasource, are trying to draw too much data or that you are over styling it (or using an inefficient SLD file).

For more details of possible things to work on see http://www.slideshare.net/mobile/geosolutions/geoserver-on-steroids

  • 1
    What do you mean by inefficient SLD file?
    – user1919
    Jun 8, 2017 at 10:04
  • @user1919, such as using line borders for your roads at multiple zoom levels. Sep 22, 2018 at 6:51

Try adding an index to your shapefile, as suggested @iant.

The command for Geoserver is:

java -cp .:gt-shapefile-2.7.3.jar org.geotools.data.shapefile.indexed.ShapeFileIndexer /path/to/your/shapefile.shp

and it must be run from the geoserver/WEB-INF/lib directory as a user who can read/write in the shapefile directory. Depending on your geoserver version you might have to change the jar version numbers.

I first saw it mentioned here where other methods for creating the index are cited.

  • I am using sql server 2008.
    – kinkajou
    Mar 26, 2012 at 4:44

Not the answer you're looking for? Browse other questions tagged or ask your own question.