Producers or consumers may be closed after the client exits out of scope

Here is the reproduction code:
```js
const Pulsar = require('pulsar-client');

(async () => {
// Create a client
const client = new Pulsar.Client({
serviceUrl: 'pulsar://localhost:6650'
});

// Create a producer
const producer = await client.createProducer({
topic: 'persistent://public/default/my-topic',
});

const sendRecords = async () => {
// Send a message
await producer.send({
data: Buffer.from("hello")
});

console.log("sent hello")
setTimeout(()=>sendRecords(), 1000)
}

await sendRecords();
})();
```

The output would be like:
```
➜ node node index.js
sent hello
sent hello
sent hello
sent hello
sent hello
sent hello
sent hello
node:internal/process/promises:288
triggerUncaughtException(err, true /* fromPromise */);
^

[Error: Failed to send message: AlreadyClosed]

Node.js v18.19.0
```
After the client exits the outer function's scope, it will eventually be garbage collected. This closes the producers and causes the AlreadyClosed issue.

A workaround is to pass the client ref to the sendRecords function:
```js
const sendRecords = async (client) => {
// Send a message
await producer.send({
data: Buffer.from("hello")
});

console.log("sent hello")
setTimeout(()=>sendRecords(client), 1000)
}

await sendRecords(client);
```

And it works.

A better approach is to keep a reference to the client inside the producer or consumer. This way, as long as we hold a reference to the producer or consumer, the client object will not be garbage collected.

No source files or tests are named. Start by running the reproduction against a local Pulsar broker, then trace producer, consumer, and client lifetime handling in the Node.js client; done means a producer or consumer keeps its client usable after the outer function scope ends and a regression test prevents AlreadyClosed during continued sends.

Written by the indexing model from the issue text.