https://github.com/damiancipolat/cluster-servers
Examples and techniques creating http servers using nodejs cluster module.
https://github.com/damiancipolat/cluster-servers
api es6 ipc nodejs parallelism pm2 server
Last synced: about 2 months ago
JSON representation
Examples and techniques creating http servers using nodejs cluster module.
- Host: GitHub
- URL: https://github.com/damiancipolat/cluster-servers
- Owner: damiancipolat
- Created: 2019-02-17T19:22:57.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2019-02-18T01:18:35.000Z (over 7 years ago)
- Last Synced: 2025-05-20T01:13:39.014Z (about 1 year ago)
- Topics: api, es6, ipc, nodejs, parallelism, pm2, server
- Language: JavaScript
- Homepage:
- Size: 265 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Cluster servers in Node.js
Examples and techniques of how to create a nodejs http server using the cluster module, creating fork process.
- **[Share ports]**
- **[Fork server]**
- **[PM2 & cluster]**
- **[Docker & cluster]**
[Share ports]:https://github.com/damiancipolat/Cluster-servers/blob/master/README.md#share-ports
[Fork server]:https://github.com/damiancipolat/Cluster-servers/blob/master/README.md#forked-server
[PM2 & cluster]:https://github.com/damiancipolat/Cluster-servers/blob/master/README.md#pm2--cluster-server
[Docker & cluster]:https://github.com/damiancipolat/Cluster-servers/blob/master/README.md#docker--cluster-server
## Share ports:
In this example I create a basic http server using the "http" module. We share the same por for all the workers, in order to share the socket to the workers.
**To run:**
```js
$ node cluset-server/cluster-server.js
```
And when we run the process the console output will be:
```sh
Master server created!!
PID: 13232 running in PORT: 8000
Worker created, PID: 13238 PORT: 8000
Worker created, PID: 13243 PORT: 8000
Worker created, PID: 13244 PORT: 8000
Worker created, PID: 13250 PORT: 8000
```
Running a ps-la, nodejs have created a different process:
```sh
F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME CMD
0 S 1000 13232 5806 0 80 0 - 219231 ep_pol pts/21 00:00:00 node
0 S 1000 13238 13232 0 80 0 - 218460 ep_pol pts/21 00:00:00 node
0 S 1000 13243 13232 0 80 0 - 218460 ep_pol pts/21 00:00:00 node
0 S 1000 13244 13232 0 80 0 - 218460 ep_pol pts/21 00:00:00 node
0 S 1000 13250 13232 0 80 0 - 218460 ep_pol pts/21 00:00:00 node
4 R 1000 13317 13265 0 80 0 - 7872 - pts/35 00:00:00 ps
```
Running lsof -i :8000, the master process PID 13232 is the only who use the port "8000":
```sh
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
node 13232 damianlinux 16u IPv6 93000 0t0 TCP *:8000 (LISTEN)
```
## Forked server:
In this code I will create a server that use fork process to create workers and communicate between them using a simple IPC.
**To run:**
```sh
$ node fork-server/app.js
```
All workers answer each request and using ipc send to the parent process the response to send at the client. In the master process there are a list of process reference. The way to delegate the answer of each request is using a random way. You can take a look of the message broker created by me in other project.
## PM2 & cluster server:
In the last examples we learn how to create our custom load balancer process using the children forked process, but in a production environment, maybe is'nt the best way. I think could be a better option work with a robust process manager. One good option is PM2 http://pm2.keymetrics.io. PM2 offer us the way to work as a load process balance for us, take a look of this link http://pm2.keymetrics.io/docs/usage/cluster-mode/.
**Configuration, define to pm2 the number of children process to create:**
```json
{
"apps" : [{
"script" : "api.js",
"instances" : "4",
"exec_mode" : "cluster"
}]
}
```
**To run:**
```sh
$ pm2 start process.json
```
**Monitoring:**
To see the list of process running:
```sh
$ pm2 list
┌──────────┬────┬─────────┬─────────┬───────┬────────┬─────────┬────────┬─────┬───────────┬─────────────┬──────────┐
│ App name │ id │ version │ mode │ pid │ status │ restart │ uptime │ cpu │ mem │ user │ watching │
├──────────┼────┼─────────┼─────────┼───────┼────────┼─────────┼────────┼─────┼───────────┼─────────────┼──────────┤
│ api │ 0 │ N/A │ cluster │ 26085 │ online │ 0 │ 2m │ 0% │ 37.3 MB │ damianlinux │ disabled │
│ api │ 1 │ N/A │ cluster │ 26091 │ online │ 0 │ 2m │ 0% │ 37.1 MB │ damianlinux │ disabled │
│ api │ 2 │ N/A │ cluster │ 26097 │ online │ 0 │ 2m │ 0% │ 36.8 MB │ damianlinux │ disabled │
│ api │ 3 │ N/A │ cluster │ 26107 │ online │ 0 │ 2m │ 0% │ 37.2 MB │ damianlinux │ disabled |
```
To see a console UI:
```sh
$ pm2 monit
```
(Left pm2 list / Right pm2 monit)
To finish all the process:
```sh
$ pm2 kill
```
Testing the process, using curl we can see that pm2 manage the request sending it to differents process, some like a load balancer. I'm not sure what is the distribution algorithm used.??
```sh
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27074
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27042
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27050
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27060
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27074
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27042
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27050
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27060
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27074
damianlinux@damianlinux:~/Escritorio/repo/Cluster-servers$ curl http://127.0.0.1:8000
Hello from worker PID:27042
```
**Note:**
Be sure your application is stateless meaning that no local data is stored in the process, for example sessions/websocket connections, session-memory and related. Use Redis, Mongo or other databases to share states between processes.
Another resource on how to write efficient, production ready stateless application is The Twelve Factor Application manifesto. https://12factor.net
## Docker & cluster server:
Todo