Sentinel-1 batch processing error SNAP (and snappy) v.8

Microwave Toolbox Problem Reports
#1

Hello everyone,

I’m doing a batch preprocessing of Sentinel-1 data (a dataset consisting of more than 100 images) using the version of snappy API configured in SNAP v.8.0. I wrote a well-functioning python script, except that doing the tests it processed the first 10 images well and then it stopped (without error messages) although still in process. Believing it was a python / snappy problem I tried directly on the SNAP software and the problem persists. Arrived at the eleventh image it gives me the error message (“some errors occured”) and the process crashed. Now I’m processing 10 images at a time, and the result are good.
Anyone can explain and / or solve this problem? I don’t think it’s a memory problem, with SNAP version 7 I worked with more images in less performing PCs.

P.S. The workflow I’m testing is: ApplyOrbit / ThermalNoiseRemoval / Calibration / TerrainFlattening / TerrainCorrection / Subset

Thanks!
Best Regards

Giandomenico De Luca

#2

I wouldn’t rule out memory issues – SNAP 7 used a different Java runtime, and memory management when using Java from Python is not simple.

There are tools for all the supported OS’s to monitor CPU and memory usage. It would be good to mention your OS details.

You need to provide more details so we can understand the problem.

When you run the workflow in SNAP look for hs_err*.log files in your home directory from the time of the error.

When using Python, you can edit snappy.ini to enable debugging by changing the line # debug: False to debug: True.