删除Google Cloud Functions的tmp文件夹中的文件

Question

I'm currently writing into files in a google cloud functions in my os.tmpdir folder. Now I know that each time the functions fires is slightly different and the tmp folder will never be the same between function calls (can't reference the same tmp folder twice). My question is whether or not I need to delete the file before the function finishes execution or if all the tmp memory is dereferenced automatically.

edit: When I write to os.tmpdir I can see the file when I look into the directory but on subsequent calls I see nothing. I assume this means the file somehow no longer exists (unlikely) or I am looking in a different place.

below is the code I user to write the file.

var wstream = fs.createWriteStream(os.tmpdir() + '/myOutput.txt');


wstream.write('Hello world!\n');
wstream.write('Another line\n');

wstream.end();      

wstream.on('finish', function () {
        console.log('file has been written');
        fs.readdir(os.tmpdir(), (err, files) => {
            console.log(files.length);
            files.forEach(file => {
                        console.log("Hey Again3", file);
                });
            })

edit 2: I've been running a few tests and the occurrence of a new system happens far more often than indicated in this question (I've observed every single time vs what I was told here that it would happen almost never). I also asked another question and received slightly conflicting advise from a different firebase employee that was more inline with the personal tests here: Can't Find file after having written to it in Google Cloud Functions

The advise stays consistent but I not once observed have the exact same file system between multiple calls of a single function as indicated below.

Answer 1

Firstly, the tmp folder identified by os.tmpdir() is always the same. If you're observing something different, please clarify exactly what you're seeing.

Secondly, you are not obliged to delete files out of tmpdir, but they occupy memory (it's a memory-based filesystem). Leaving those files around makes your future function invocations more expensive, since you are paying for memory over time. Also, you may risk running out of memory as you accumulate temp files. To be most efficient, you should always delete any files you create before the function finishes.