Unhandled Promise Rejection On Failed createProducer

[savearray2]

When creating a producer using a client URL that is intentionally incorrect via the following:

const client = new Pulsar.Client({
	serviceUrl: 'pulsar://this-will-fail:6650',
	ioThreads: 6
})
client.createProducer({
	topic: 'persistent://test/a/b',
}).then((producer) => {
	// ...
}).catch((err) => {
	// ...
})

I get the following unhandled promise rejection warning:

[ClientConnection:469] [<none> -> pulsar://this-will-fail:6650] Resolve error: asio.netdb:1 : Host not found (authoritative)
(node:15) UnhandledPromiseRejectionWarning: Error: Failed to create producer: ConnectError
(node:15) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). To terminate the node process on unhandled promise rejection, use the CLI flag `--unhandled-rejections=strict` (see https://nodejs.org/api/cli.html#cli_unhandled_rejections_mode). (rejection id: 1)
(node:15) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.

Upon further inspection of the code the error originates from the following line:

pulsar-client-node/src/Consumer.cc

Line 134 in eaba0dc

SetError(std::string("Failed to create consumer: ") + pulsar_result_str(result));

Where ConsumerNewInstanceWorker is a Napi::AsyncWorker:

pulsar-client-node/src/Consumer.cc

Lines 94 to 102 in eaba0dc

	class ConsumerNewInstanceWorker : public Napi::AsyncWorker {
	public:
	ConsumerNewInstanceWorker(const Napi::Promise::Deferred &deferred, pulsar_client_t *cClient,
	ConsumerConfig *consumerConfig, std::shared_ptr<CConsumerWrapper> consumerWrapper)
	: AsyncWorker(Napi::Function::New(deferred.Promise().Env(), [](const Napi::CallbackInfo &info) {})),
	deferred(deferred),
	cClient(cClient),
	consumerConfig(consumerConfig),
	consumerWrapper(consumerWrapper) {}

I did some quick checking into using Promises with AsyncWorker and there is a proposed solution mentioned here on the NAPI repository: nodejs/node-addon-api#231 (comment)

I'd be willing to incorporate this solution and create a pull request, if it would be welcome, but I wanted to discuss it here first before making any changes.

I'd also be willing to help out with some of the other requests here, as well.
I'll be using this in production going forward.

[savearray2]

Well, this is strange.
I cloned the project and created a simple js file to reproduce the error:

const Pulsar = require('.')
const client = new Pulsar.Client({
	serviceUrl: 'pulsar://this-will-fail:6650'
})
client.createProducer({
	topic: 'persistent://a/b/c',
}).then(() => {
	console.log('success')
}).catch((err) => {
	console.log('normal err')
	console.log(err)
})

This leads to the following output:

# node test_me.js
2020-09-11 16:39:03.046 INFO  [0x700008f87000] ConnectionPool:85 | Created connection for pulsar://this-will-fail:6650
2020-09-11 16:39:03.058 ERROR [0x70000a813000] ClientConnection:469 | [<none> -> pulsar://this-will-fail:6650] Resolve error: asio.netdb:1 : Host not found (authoritative)
2020-09-11 16:39:03.058 INFO  [0x70000a813000] ClientConnection:1387 | [<none> -> pulsar://this-will-fail:6650] Connection closed
2020-09-11 16:39:03.058 ERROR [0x70000a813000] ClientImpl:181 | Error Checking/Getting Partition Metadata while creating producer on persistent://a/b/c -- ConnectError
2020-09-11 16:39:03.058 INFO  [0x70000a813000] ClientConnection:238 | [<none> -> pulsar://this-will-fail:6650] Destroyed connection
normal err
[Error: Failed to create producer: ConnectError]
[1]    45767 segmentation fault  node test_me.js

This still seems to be related to the comment in the NAPI repository mentioned above. According to the comment, a change needed to be made so that exceptions do not occur on exit.

I'd like to get some input from you guys on how to proceed.

[savearray2]

Okay, I probably should have done some more research (and gotten more sleep) before posting this. Sorry about this mess.
The problem with the unhandled promise rejection is not with the pulsar library. I had a typo in my testing framework code.

The second problem with the segmentation fault happens when the client is not closed properly before exit. It may be desirable to include code in the library to attach to a process exit to automatically close out the client if it hasn't been already, but this might add unneeded complexity.

The following code fixes the test:

const Pulsar = require('.')
const client = new Pulsar.Client({
	serviceUrl: 'pulsar://this-will-fail:6650'
})
client.createProducer({
	topic: 'persistent://a/b/c',
}).then(() => {
	console.log('success')
	client.close()
}).catch((err) => {
	console.log('normal err')
	console.log(err)
	client.close()
})

Please note the inclusion of client.close().

This issue may be closed at the maintainers' leisure.
Again, I apologize for the trouble.

[sijie]

@savearray2 thank you for your explanation! Close this issue since it seems that the problem has been resolved.