Open offtherailz opened 10 years ago
To avoid this problem, we found a workaround: Add this 2 lines to your config.yaml
globalParallelFetches: 1
perHostParallelFetches: 1
basically it tells to MapFish to do not create the thread pool. Obviously this change will slow down the pdf creation, but can avoid the jvm out of memory problem.
The out of memory error like this:
The GeoServer log notifies:
In the standard output of your jvm you can also find some errors like:
Basically the MapFish print and the related plugin of GeoServer create a thread pool for every pdf generation request, and it doesn't dispose these threads after the print process end.