Zombie Process'

[DaneEveritt]

It appears that there is a bug with pty.js (see here and here) which prevents stopped processes from actually closing out correctly, so they become zombies to the scales.js process.

Not sure how we will go about fixing this, but at this point it doesn't look like there is too much we can do unfortunately. It was suggested to use waitpid(), however that doesn't seem to be compatible with NodeJS 0.12, unless that was just a bug on my system.

To reproduce this, simply start a server and then stop it. You can do this over and over, and then run ps aux | grep <process> (i.e. java) and you will see a long list of zombie processes form over time.

[DDynamic]

@LordRalex do you think you'll be able to fix this? Or will we have to wait for @DaneEveritt to return? This is a pretty big issue.

[LordRalex]

@BlameDylan I have 0 knowledge of nodejs or Scales. I can BS a fix if I spend time on it, but at this point, that time is hard to get. I'd rather have someone help patch it and submit a PR, since that requires less time for me to allocate.

[DDynamic]

I'm currently learn NodeJS. Maybe I can do something.

[Mstrodl]

Is there any sort of work-around for this? Is it as simple as restarting Scales?

[aadeg]

It will be fixed soon?

[DDynamic]

@aadeg I think it has been fixed in the "zombies" branch.

[aadeg]

@BlameDylan The last commit was 1 month ago... I hope it will be fixed.

[Mstrodl]

@aadeg The commit message says

Attempt to fix zombie processes

[DaneEveritt]

You can fix the Zombie Process issue by just restarting Scales for now. There is nothing we can do until it is fixed in an upstream dependency branch.