e2e test => Jest did not exit one second after the test run has completed.

[danieldaeschle]

I'm submitting a...

[ ] Regression 
[X] Bug report
[ ] Feature request
[ ] Documentation issue or request
[ ] Support request => Please do not submit support request here, instead post your question on Stack Overflow.

Current behavior

Run jest with --runInBand parameter.

Expected behavior

If i run it without this parameter it works like expected.
But on gitlab CI i don't use this parameter and it doesn't work too.

Minimal reproduction of the problem with instructions

nestjs-jest-e2e-test-bug.tar.gz

Install redis server locally and run this commands:

npm install
npm run test:e2e -- --runInBand

What is the motivation / use case for changing the behavior?

It doesn't work and it should work because this is a nice framework!

Environment

Nest version: `^5.4.0`
 
For Tooling issues:
- Node version: 10.15
- Platform:  Linux

[danieldaeschle]

Just tested it with the newest version of nest (5.7.0) which also doesn't work.

[kamilmysliwiec]

We can't help you without at least a small repository which reproduces your issue.

[danieldaeschle]

@kamilmysliwiec i updated the Minimal reproduction section.

[immanuel192]

confirm that v5.6.2 also got this issue. I have to use --forceExit to force jest terminate

[danieldaeschle]

@kamilmysliwiec did you found something?

[danieldaeschle]

push

[kamilmysliwiec]

This is not a framework bug - you have an issue in your code. If you remove the redis part, it's going to work correctly which means that some of your async operations (or even connection) haven't fulfilled correctly.

Please, use StackOverflow for such questions.

[danieldaeschle]

I already started a stack overflow question. But i got no answer.
I'm not the only one with this problem.

If you really think that my code is wrong, then tell me please where the wrong code is.

[ghost]

@danieldaeschle I ran into a similar issue, and determined what my issue was.
Instead of redis, I was using socket.io-client to connect to a Nestjs Gateway that I was testing.

When using socket.io-client, I establish a connection by calling a .connect(..) method and providing the host and port of where my server is listening on.

It turns out, I was accidentally calling io.connect('localhost:3000') in more than one location in my test file. (Once in a beforeEach section, and once within the unit test itself).

To clean up my connections, I had to call a .disconnect(true) function. But before this worked for me, I had to catch the duplicate .connect(..) call that I was making, which likely meant there was some unhandled async operation going on that I wasn't cleaning up before jest finished.

All in all I'm really glad Jest actually doesn't kill itself when it notices unhandled async ops.
I'm also glad I didn't use "--forceExit" because that would have put a band-aid over the fact that I wasn't cleaning up my connections properly instead of letting me find the root cause.

@kamilmysliwiec 's advice that your problem isn't a framework bug and that the issue is likely in your code is probably accurate.

Go ahead and find out how people close redis connections, then verify you are closing any connections you've opened in the right location (in the end of your unit test, in an afterEach block, or in an afterAll block.), and also verify that you aren't opening the connection more than once accidentally like I was.

[gjermundnor]

I had the same issue.

I was connecting to Pusher with a real key. When switching to a fake key, it solved the problem for me.

[DaGaMs]

I have this issue and I do think there's something wrong on the framework level, or at least the documentation is unclear. In my case, I try to run a websocket server with the socket.io-redis adapter. I am not actually doing anything with redis yet (ie no rooms are created, no emits triggered). As soon as I switch over from the standard adapter to the redis-io adapter, the Jest tests hang. Here's the relevant setup part of my tests:

beforeAll(async () => {
    const moduleFixture: TestingModule = await Test.createTestingModule({
      imports: [AppModule],
    }).compile();

    app = moduleFixture.createNestApplication();
    app.enableCors();

    // IF I UNCOMMENT THIS LINE, JEST FINISHES FINE
    app.useWebSocketAdapter(new RedisIoAdapter(app));

    await app.init();

    const address = app
      .getHttpServer()
      .listen()
      .address();
    baseAddress = `http://[${address.address}]:${address.port}/editing/story`;
  });
  afterAll(async () => {
    await app.close();
  });

RedisIOAdapter is copy-pasted from https://docs.nestjs.com/websockets/adapter. Redis is running happily on the machine. I am 100% sure i'm closing all connections in the test (in afterEach). Any idea what might be going on?

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.