Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/mikey-t/node-ssh-reboot-test


https://github.com/mikey-t/node-ssh-reboot-test

Last synced: about 1 month ago
JSON representation

Awesome Lists containing this project

README 

        
# node-ssh-reboot-test



Attempting to repro https://github.com/steelbrain/node-ssh/issues/421.



## Description



When using the npm library `node-ssh` and executing a reboot command (i.e. `sudo reboot`), an error occurs but is not caught. The node process crashes with an error like this:



```

Error: read ECONNRESET

    at TCP.onStreamRead (node:internal/stream_base_commons:217:20) {

  errno: -4077,

  code: 'ECONNRESET',

  syscall: 'read',

  level: 'client-socket'

}

```



For me this only happens sometimes - I haven't found a pattern - it seems to happen maybe 1 out of every 4 or 5 times.



## Notes on Variables



Stuff I'm using (see package.json for versions):



- Node 18.18.0

- Typescript

- ts-node (with swc as the compiler - see tsconfig.json)

- Accessing an Ubuntu 20 instance on AWS Lightsail

- Running on windows in powershell 7



## Setup and Run Test



- Run: `npm install`

- Copy `.env.template` to `.env` and update values



Run the test from @steelbrain:

```

npm run index

```

Things I changed in steelbrain's test:

- I added `.then().catch()` on main

- I'm connecting with privateKeyPath instead of password



I created another test with the error event listener commented out:



```

npm run another

```



## Results



I get different results for both of these tests when running them multiple times. For the "index" test I sometimes get the expected result:



```

{ code: 0, signal: null, stdout: 'hi', stderr: '' }

{ code: null, signal: null, stdout: '', stderr: '' }

no uncaught errors - exiting

```



But other times I get this:



```

{ code: 0, signal: null, stdout: 'hi', stderr: '' }

error Error: read ECONNRESET

    at TCP.onStreamRead (node:internal/stream_base_commons:217:20) {

  errno: -4077,

  code: 'ECONNRESET',

  syscall: 'read',

  level: 'client-socket'

}

{ code: null, signal: null, stdout: '', stderr: '' }

no uncaught errors - exiting

```



Note that it didn't crash the node process.



When I run the "another" test I sometimes get no errors:



```

{ code: 0, signal: null, stdout: 'hi', stderr: '' }

{ code: null, signal: null, stdout: '', stderr: '' }

no uncaught errors - exiting

```



But other times I get this:



```

{ code: 0, signal: null, stdout: 'hi', stderr: '' }

Error: read ECONNRESET

    at TCP.onStreamRead (node:internal/stream_base_commons:217:20) {

  errno: -4077,

  code: 'ECONNRESET',

  syscall: 'read',

  level: 'client-socket'

}

```



This is before I added the `process.on` listeners for `unhandledRejection` and `uncaughtException`. Even without those I would have expected at least the `catch` block to be hit, but it wasn't.



After adding the `process.on` listeners I tried 5 or 6 more times and wasn't getting any errors, but then finally got this:



```

{ code: 0, signal: null, stdout: 'hi', stderr: '' }

Caught exception: Error: read ECONNRESET

Exception origin: uncaughtException

```

Note that this time the `.then().catch()` is never hit, but the `process.on` event for `uncaughtException` IS hit. This doesn't seem like it should be expected. Omitting a handler for `ssh.connection?.on` event `error` shouldn't cause try/catch to fail.



## Additional Notes



I cloned this test repo into WSL and tried reproducing the issue but was not able to. I guess my next step will be to build the node-ssh lib on windows, link it and try fiddling with this bit of code since it seems like there's some sort of race condition around the error emitting event, the removal of the event listener and the "normal" rejection.