foreman exits immediately

[tkral]

foreman exits immediately

[tkral]

Turns out this was nothing foreman related. There was something wrong with my system. Never figured it out, but a reboot solved the problem.

[ddollar]

What does your Procfile look like?

[tkral]

I'm trying to use the Play framework so my Procfile is just this single line:

web: /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader --http.port=$PORT

[ddollar]

Does running this command verbatim in your console work?

env PORT=5000 /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader --http.port=$PORT

[tkral]

So that command doesn't work:

tkral@tkral-wsl:~$ env PORT=5000 /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader --http.port=$PORT
...
Listening for transport dt_socket at address: 8000
14:43:16,473 INFO ~ Starting /home/tkral/dev/samples/dataloader
14:43:16,907 WARN ~ You're running Play! in DEV mode
Exception in thread "main" java.lang.NumberFormatException: For input string: ""
at java.lang.NumberFormatException.forInputString(NumberFormatException.java:48)
at java.lang.Integer.parseInt(Integer.java:470)
at java.lang.Integer.parseInt(Integer.java:499)
at play.server.Server.(Server.java:33)
at play.server.Server.main(Server.java:160)

However, that seems to be a red herring. If I remove the --http.port option so that my Procfile looks like this:

web: /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader

Then the console command works:

tkral@tkral-wsl:~$ env PORT=5000 /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader
...
Listening for transport dt_socket at address: 8000
14:48:12,628 INFO ~ Starting /home/tkral/dev/samples/dataloader
14:48:12,980 WARN ~ You're running Play! in DEV mode
14:48:13,031 INFO ~ Listening for HTTP on port 9000 (Waiting a first request to start) ...
14:48:16,260 INFO ~ Connected to jdbc:h2:/home/tkral/dev/samples/dataloader/db/h2/play;MODE=MYSQL;DB_CLOSE_ON_EXIT=FALSE
14:48:16,757 INFO ~ Application 'dataloader' is now started !

But foreman is still exiting immediately:

tkral@tkral-wsl:/dev/samples/dataloader[develop_]$ cat Procfile
web: /home/tkral/dev/play/play run /home/tkral/dev/samples/dataloader
tkral@tkral-wsl:/dev/samples/dataloader[develop_]$ foreman start
14:48:55 web.1 | started with pid 6963
14:48:55 web.1 | process exiting
14:48:55 web.1 | process terminated
14:48:55 system | sending SIGTERM to all processes

[ddollar]

If you run the command copied verbatim out of the Procfile, does it work at your terminal?

It seems like the program is terminating immediately with no output when running under foreman.

[fdr]

I have also encountered this issue just now (start a process, terminate immediately, no output, although my Procfile is nothing special. I am using the standalone foreman. It is not reproducing with the rvm+gem installed one I usually use:

$ bundle exec foreman start -f /tmp/Procfile  
23:52:32 web.1     | started with pid 17721
23:52:33 web.1     | hello

vs

./foreman/foreman start -f /tmp/Procfile 
23:54:23 web.1     | started with pid 17909
23:54:24 web.1     | process exiting
23:54:24 web.1     | process terminated
23:54:24 system    | sending SIGTERM to all processes

Trivial Procfile and program:

web: bash /tmp/blah $PORT

echo hello
sleep 10000

[1xch]

I'm getting this...everything gets shut down immediately(foreman start and bundle exec foreman start. The processes work fine from the command line.

14:20:53 beanstalkd.1 | started with pid 8477
14:20:53 stalker.1 | started with pid 8480
14:20:53 beanstalkd.1 | process exiting
14:20:53 stalker.1 | process exiting
14:20:53 stalker.1 | process terminated
14:20:53 system | sending SIGTERM to all processes

[noahd1]

I'm also seeing this issue.

$ foreman start -f features/support/Procfile.dev -p 4001
14:56:59 calculator.1         | started with pid 1589
14:56:59 calculator_resque.1  | started with pid 1590
14:56:59 calculator_resque.1  | process terminated
14:56:59 system               | sending SIGTERM to all processes
$ 

Running the two commands in my Procfile without foreman work fine.

[noahd1]

The problem (above) I experienced with the latest version (0.30.1). I was able to avoid the issue by installing an older version of foreman. Specifically, 0.26.1 and 0.27.0 seem to work fine. Gems after these did not work for me, for this reason and other reasons (some just blew up immediately on calling foreman).

[gnufied]

@noahd1 Yup, agreed. moving to 0.26.1 fixes the problem and latest versions seem to have this problem. Can this ticket be reopened, since this is valid one:

16:06:51 xxxx.1     | started with pid 11846
16:06:51 yyyy.1      | started with pid 11849
16:06:51 zzzzz.1  | started with pid 11848
16:06:51 ccccc.1  | started with pid 11847
16:06:51 ddddd.1     | started with pid 11851
16:06:51 ddddd.1     | process terminated
16:06:51 system       | sending SIGTERM to all processes
16:06:51 system       | sending SIGTERM to pid 11847
16:06:51 system       | sending SIGTERM to pid 11848
16:06:51 system       | sending SIGTERM to pid 11849
16:06:51 system       | sending SIGTERM to pid 11846
16:06:51 ccccc.1      | process terminated
16:06:51 zzzzz.1  | process terminated
16:06:51 yyyy.1  | process terminated
16:06:51 xxxx.1     | process terminated

[ehaughee]

I am having a similar issue on the latest and on 0.26.1. One of the processes might be daemonizing itself (rake thinking_sphinx:start) but the other doesn't seem to do so (rackup private_pub.ru -s thin -E production). Also, I am running Rails 3.2 on OS X 10.7. Let me know if any more background information is necessary.

Here is my output:

$ foreman start            
12:25:43 private_pub.1      | started with pid 9129
12:25:43 thinking_sphinx.1  | started with pid 9130
12:25:50 thinking_sphinx.1  | Started successfully (pid 9139).
12:25:50 thinking_sphinx.1  | process terminated
12:25:50 system             | sending SIGTERM to all processes
12:25:50 system             | sending SIGTERM to pid 9129
12:25:51 private_pub.1      | >> Thin web server (v1.3.1 codename Triple Espresso)
12:25:51 private_pub.1      | >> Maximum connections set to 1024
12:25:51 private_pub.1      | >> Listening on 0.0.0.0:9292, CTRL+C to stop
12:25:51 private_pub.1      | process terminated

And my Procfile:

private_pub:        rackup private_pub.ru -s thin -E production
thinking_sphinx:    rake thinking_sphinx:start

Both of these commands run when run verbatim in Bash. As I mentioned, I'm pretty sure thinking_sphinx daemonizes itself but I'm also pretty sure private_pub/faye does not.

##EDIT:
I don't know where the issue actually lies, but the following blog post got me up and running. Thinking_sphinx was causing issues with foreman: http://icelab.com.au/articles/run-your-own-piece-of-heroku-with-foreman/

[nikhilbhavsar5895]

Hi i am also getting same issue
14:25:38 server.1 | started with pid 14836
14:25:38 frontend.1 | started with pid 14837
14:25:38 server.1 | 2020/08/12 14:25:38 mmap: bad file descriptor
14:25:38 server.1 | exited with code 1
14:25:38 system | sending SIGTERM to all processes
14:25:38 frontend.1 | terminated by SIGTERM

My procfile contains
server: bee run -downdoc=true -gendoc=true -e=frontend -e=conf -e=logs -e=datastructs -e=filters -e=grifts -e=out -e=inp -e=database -e=events -e=tasks -e=templates
frontend: cd frontend && npm run dev

[tzipisheinin]

It seems like the program is terminating immediately with no output when running under foreman.

Is this issue still happening?
My Procfile.https look like this:
web: thin start --ssl --ssl-key-file server.key --ssl-cert-file server.crt --port 3000 redis: redis-server worker: bundle exec rake resque:work QUEUE='*' scheduler: bundle exec rake resque:scheduler guard: guard -P livereload
and I just find out that running the first command - web using forman start -f Procfile.https
doesn't show the errors of this command, but when I run this line alone in terminal - I can see the exact errors.
Does foreman hiding the errors, or maybe he doesn't even get those errors from his child commands

[rakaur]

Turns out my server wasn't starting because of some weird file in the way, and the error wasn't being shown anywhere.

Still happens to me. Makes using bin/dev or foreman impossible in Docker. I can run the container and run bin/dev or foreman start -f Procfile.dev in a shell in the container and it works fine. The minute I try to make it the entrypoint/cmd, it exits immediately.

streams_1   | 16:29:59 web.1  | started with pid 10
streams_1   | 16:29:59 js.1   | started with pid 11
streams_1   | 16:29:59 css.1  | started with pid 14
streams_1   | 16:30:00 css.1  | yarn run v1.22.17
streams_1   | 16:30:00 js.1   | yarn run v1.22.17
streams_1   | 16:30:00 css.1  | $ sass ./app/assets/stylesheets/application.bulma.scss ./app/assets/builds/application.css --no-source-map --load-path=node_modules --watch
streams_1   | 16:30:00 js.1   | $ esbuild app/javascript/*.* --bundle --sourcemap --outdir=app/assets/builds --watch
streams_1   | 16:30:00 js.1   | Done in 0.26s.
streams_1   | 16:30:00 js.1   | exited with code 0
streams_1   | 16:30:00 system | sending SIGTERM to all processes
streams_1   | 16:30:00 web.1  | terminated by SIGTERM
streams_1   | 16:30:01 css.1  | exited with code 1
streams_streams_1 exited with code 0