I’ve built a python script for batch processing sentinel2 products. It works fine except that memory consumption starts growing until it reaches the limit allowed to java and it stops working.
I’ve searched the forums and it seems this issue is not new. The only fix I’ve found is this:
However, I was wondering if there’re any further developments since this was discussed for the last time, around 2019.
Also, another user claims memory issues are related to GPF.createProduct. Is that correct?
in the past we have worked more on the java side to improve performance with respect to memory and speed. Probably python will benefit from it too, but we haven’t validated this.
That product.dispose() is not effective on products created with GPF.createProduct() I haven’t noticed yet. But also not investigated.
I’ve created ticket for this issue, so we can follow up on this.