GridFS.put() performs a read on 'data' which is a string of bytes or a file-like object with a read() method. For a file whose size is larger than the amount of available system memory this can create a out-of-memory kill from the operating system.

I've forked the python driver and modified put() to accept a keyword 'huge' that when presented with a file-like object will read 'chunk_size' bytes from the file-object at a time (with chunk_size gotten from the GridIn object).