Can't kill zombie process

[willgdjones]

Running heroku local fails with an error.

However, my app mysteriously still runs on 0.0.0.0:5000, and leaves zombies processes that I can't kill.

When I try to restart heroku local it tells me the socket is already in use.

Can you give me any suggestions as to how this could have happened?

[lastobelus]

I have a similar issue. If I run heroku local, when I kill it (ex: with ctrl-c in the terminal where it is running) it often leaves zombie one or more of the processes specified in the Procfile.

[lastobelus]

A little more testing: only the web process gets orphaned (not a zombie), and it happens every time.

Here is my Procfile:

web: bundle exec rails s
webpacker: ./bin/webpack-dev-server

for Rails 5.1 app.

Every time I quit heroku local it leaves a ruby process running owned by a node process. I guess they're not zombies but the node process is an orphan, and owns the ruby process.

So every time I need to restart rails locally, I have to look for both the ruby & node processes & kill them. However I can't kill them with kill or kill -9 I have to use Activity Monitor and force quit them.

[jdx]

@lastobelus are you on the latest version (6.13.13)? I think we resolved this issue

[yoeriboven]

Same problem here. Local server still runs after stopping it with Ctrl-C. Only a web: process in the Procfile.

The others are shown with the ps aux |grep php-fpm command. I'm on heroku-cli/6.14.13-6ab5022 (darwin-x64) node-v8.4.0.

[lastobelus]

not happening any more for me on latest version.

[marckohlbrugge]

I'm experiencing the same issue on heroku-cli/6.16.3-4239951 (darwin-x64) node-v9.9.0