Thanks @UnborN for your suggestion.
In case to enable multiplequery I should modify code directly inside contrib-node code, is it correct?
Thanks @TotallyInformation for your suggestion.
I already tried to retrieve data with Python code (pythonshell node) but I still have same problem (worse: python script execution consume also a lot of RAM).
I already tried to save data in a file and to read it (file in node), but I still have the same starvation problem: read this big file blocks node-red (also if the problem is visible for bigger query respect to use mysql directly).
I'll follow your indications and try to break data into chunks (there is an indication also in Exec and daemon nodes break up a data stream), using exec node and stdout or mysql node with single query (I have to try).
Thanks @Colin for your suggestion.
I'm sorry for the question, where I could find "myself log"?