Error when running kong/httpbin:0.2.0

[MattiasGees]

Hi,

The latest release of httpbin broke the image. When I now try to start running it I get the following error:

gunicorn": executable file not found in $PATH: unknown

[pmalek]

I have tested and both 0.1.0 and 0.2.0 run without errors:

docker run -it --rm kong/httpbin:0.1.0
Unable to find image 'kong/httpbin:0.1.0' locally
0.1.0: Pulling from kong/httpbin
11e23ac719b3: Pull complete
48f131cc60da: Pull complete
a6f08b48d25e: Pull complete
4f4fb700ef54: Pull complete
1f743af032a3: Pull complete
2124c491461a: Pull complete
7042a0ef4f5e: Pull complete
Digest: sha256:b9bfd44e6aca835b6622fc0ebb03372792ac3fb5c36477c3c970691a7e30bb20
Status: Downloaded newer image for kong/httpbin:0.1.0
[2024-09-24 14:21:25 +0000] [1] [INFO] Starting gunicorn 19.9.0
[2024-09-24 14:21:25 +0000] [1] [INFO] Listening at: http://0.0.0.0:80 (1)
[2024-09-24 14:21:25 +0000] [1] [INFO] Using worker: gevent

docker run -it --rm kong/httpbin:0.2.0
Unable to find image 'kong/httpbin:0.2.0' locally
0.2.0: Pulling from kong/httpbin
e63ce922f022: Pull complete
35e7cfe3f2b2: Pull complete
50ec409c3c54: Pull complete
ce617194b43b: Pull complete
af2ee96c978a: Pull complete
Digest: sha256:c0c44152d0c9466915aaea170fd4dbe7a418829d7c479520407e4f2f9048a9aa
Status: Downloaded newer image for kong/httpbin:0.2.0
[2024-09-24 14:21:47 +0000] [1] [INFO] Starting gunicorn 21.2.0
[2024-09-24 14:21:47 +0000] [1] [INFO] Listening at: http://0.0.0.0:80 (1)
[2024-09-24 14:21:47 +0000] [1] [INFO] Using worker: gevent
[2024-09-24 14:21:47 +0000] [8] [INFO] Booting worker with pid: 8
^C[2024-09-24 14:21:53 +0000] [1] [INFO] Handling signal: int
[2024-09-24 14:21:53 +0000] [8] [INFO] Worker exiting (pid: 8)
[2024-09-24 14:21:53 +0000] [1] [INFO] Shutting down: Master

Can you share steps to reproduce?

[taer]

I have the same thing actually. I was running with the config from istio

spec:
  containers:
  - command:
    - gunicorn
    - -b
    - 0.0.0.0:8080
    - httpbin:app
    - -k
    - gevent
    env:
    - name: WORKON_HOME
      value: /tmp
    image: docker.io/kong/httpbin
    imagePullPolicy: IfNotPresent
    name: httpbin

[taer]

Just removed the command section from the launch, and here are the logs

Creating a virtualenv for this project
Pipfile: /httpbin/Pipfile
Using default python from /usr/bin/python33.10.12 to create virtualenv...
created virtual environment CPython3.10.12.final.0-64 in 204ms
  creator CPython3Posix(dest=/tmp/httpbin-30d5U6R4, clear=False, 
no_vcs_ignore=False, global=False)
  seeder FromAppData(download=False, pip=bundle, setuptools=bundle, 
wheel=bundle, via=copy, app_data_dir=/root/.local/share/virtualenv)
    added seed packages: pip==24.2, setuptools==75.1.0, wheel==0.44.0
  activators 
BashActivator,CShellActivator,FishActivator,NushellActivator,PowerShellActivator
,PythonActivator

✔ Successfully created virtual environment!
Virtualenv location: /tmp/httpbin-30d5U6R4
Error: the command gunicorn could not be found within PATH or Pipfile's [scripts].

with 0.1.0

    spec:
      containers:
      - env:
        - name: WORKON_HOME
          value: /tmp
        image: docker.io/kong/httpbin:0.1.0
        imagePullPolicy: IfNotPresent
        name: httpbin

the logs are

[2024-09-24 20:24:54 +0000] [1] [INFO] Starting gunicorn 19.9.0
[2024-09-24 20:24:54 +0000] [1] [INFO] Listening at: http://0.0.0.0:80 (1)
[2024-09-24 20:24:54 +0000] [1] [INFO] Using worker: gevent
/usr/lib/python3.8/os.py:1023: RuntimeWarning: line buffering (buffering=1) isn't supported in binary mode, the default buffer size will be used
  return io.open(fd, *args, **kwargs)
[2024-09-24 20:24:54 +0000] [9] [INFO] Booting worker with pid: 9

[taer]

the command overrides were there to change the port to 8080

[taer]

I found it. I'm not sure why this is here(again, this is the deployment yaml from Istio), but this env var breaks it WORKON_HOME=/tmp. The comment next to it says

            # Tells pipenv to use a writable directory instead of $HOME

docker run -e WORKON_HOME=/tmp --rm docker.io/kong/httpbin:0.2.0 <--- fails
docker run --rm docker.io/kong/httpbin:0.2.0 <--- works

[demikl]

Same here, starting the container as non-root in Kubernetes :

apiVersion: apps/v1
kind: Deployment
spec:
  template:
    spec:
      containers:
      - image: all-docker.artifact.cloud.socrate.vsct.fr/kong/httpbin
        name: httpbin
        securityContext:
          runAsUser: 1000

Traceback (most recent call last):
  File "/usr/local/bin/pipenv", line 8, in <module>
    sys.exit(cli())
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 1157, in __call__
    return self.main(*args, **kwargs)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/cli/options.py", line 52, in main
    return super().main(*args, **kwargs, windows_expand_args=False)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 1078, in main
    rv = self.invoke(ctx)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 1688, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 1434, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 783, in invoke
    return __callback(*args, **kwargs)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/decorators.py", line 92, in new_func
    return ctx.invoke(f, obj, *args, **kwargs)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/vendor/click/core.py", line 783, in invoke
    return __callback(*args, **kwargs)
  File "/usr/local/lib/python3.10/dist-packages/pipenv/cli/command.py", line 422, in run
    do_run(
  File "/usr/local/lib/python3.10/dist-packages/pipenv/routines/shell.py", line 68, in do_run
    ensure_project(
  File "/usr/local/lib/python3.10/dist-packages/pipenv/utils/project.py", line 40, in ensure_project
    if project.s.PIPENV_USE_SYSTEM or project.virtualenv_exists:
  File "/usr/local/lib/python3.10/dist-packages/pipenv/project.py", line 418, in virtualenv_exists
    if os.path.exists(self.virtualenv_location):
  File "/usr/local/lib/python3.10/dist-packages/pipenv/project.py", line 606, in virtualenv_location
    self._virtualenv_location = self.get_location_for_virtualenv()
  File "/usr/local/lib/python3.10/dist-packages/pipenv/project.py", line 440, in get_location_for_virtualenv
    return str(get_workon_home().joinpath(self.virtualenv_name))
  File "/usr/local/lib/python3.10/dist-packages/pipenv/utils/shell.py", line 196, in get_workon_home
    os.makedirs(expanded_path, exist_ok=True)
  File "/usr/lib/python3.10/os.py", line 215, in makedirs
    makedirs(head, exist_ok=exist_ok)
  File "/usr/lib/python3.10/os.py", line 215, in makedirs
    makedirs(head, exist_ok=exist_ok)
  File "/usr/lib/python3.10/os.py", line 225, in makedirs
    mkdir(name, mode)
PermissionError: [Errno 13] Permission denied: '/.local'
Stream closed EOF for httpbin/httpbin-76cf96d9fd-mmd6c (httpbin)

Adding a writeable mount in /.local/ solves this:

apiVersion: apps/v1
kind: Deployment
spec:
  template:
    spec:
      containers:
      - image: all-docker.artifact.cloud.socrate.vsct.fr/kong/httpbin
        name: httpbin
        securityContext:
          runAsUser: 1000
        volumeMounts:
        - mountPath: /.local
          name: venv
      volumes:
      - emptyDir: {}
        name: venv

Edit: no, it doesn't work, since this creates a new virtualenv that is not fully initialized. The image is expecting to use the already provided venv that is located at:

/root/.local/share/virtualenvs/httpbin-30d5U6R4/

[pmalek]

@demikl I've just tested and not specifying the user works well:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: httpbin
spec:
  replicas: 1
  selector:
    matchLabels:
      app: httpbin
  template:
    metadata:
      labels:
        app: httpbin
    spec:
      containers:
      - image: kong/httpbin:0.2.0
        name: httpbin

If you feel this can be improved to accommodate specifying the user please feel free to submit a PR. Having said that please bear in mind that this is a different issue than the one we're discussing here so if you feel you'd like to take this forward consider creating a separate one.

[VBranimir]

@demikl I've just tested and not specifying the user works well:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: httpbin
spec:
  replicas: 1
  selector:
    matchLabels:
      app: httpbin
  template:
    metadata:
      labels:
        app: httpbin
    spec:
      containers:
      - image: kong/httpbin:0.2.0
        name: httpbin

If you feel this can be improved to accommodate specifying the user please feel free to submit a PR. Having said that please bear in mind that this is a different issue than the one we're discussing here so if you feel you'd like to take this forward consider creating a separate one.

Reproduced problems from issue 60 and issue 62 and submitted a PR that enables usage of runAsUser :
#63

BR, Branimir

[MarkusFlorian79]

I have the same problem running the image from docker.io.

docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: exec: "gunicorn": executable file not found in $PATH: unknown

Running docker inspect on that image shows me:
"Cmd": [ "gunicorn", "-b", "0.0.0.0:8080", "httpbin:app", "-k", "gevent" ],
Running gunicorn manually in the image fails as well. While running "pipenv run gunicorn ..." works.

Building the image locally and running docker describe shows me:
"Cmd": [ "pipenv", "run", "gunicorn", "-b", "0.0.0.0:80", "httpbin:app", "-k", "gevent" ],

Is it possible that the image on docker.io was build with the old Dockerfile using the cmd without pipenv?

[pmalek]

Hi @MarkusFlorian79

I've just checked and 0.2.1 works well for me:

docker run -it --rm kong/httpbin:0.2.1
[2024-10-31 20:24:36 +0000] [1] [INFO] Starting gunicorn 21.2.0
[2024-10-31 20:24:36 +0000] [1] [INFO] Listening at: http://0.0.0.0:80 (1)
[2024-10-31 20:24:36 +0000] [1] [INFO] Using worker: gevent
[2024-10-31 20:24:36 +0000] [8] [INFO] Booting worker with pid: 8

The Cmd is set to what you described you also got when building locally.

docker inspect kong/httpbin:0.2.1 | jq '.[0].Config.Cmd'
[
  "pipenv",
  "run",
  "gunicorn",
  "-b",
  "0.0.0.0:80",
  "httpbin:app",
  "-k",
  "gevent"
]

[pmalek]

@VBranimir Can we close this one as #63 was merged?

[VBranimir]

Yes, to the best of my knowledge we can close this issue. On my side the new image from docker-hub works as expected with non-root/custom user, as well as with root. Tested both on kubernetes with "runAsUser" flag and docker with "-u" flag.

tags: "latest" and "0.2.1"
digest: sha256:5aee7b0cfa825e1317ad795f65e88f63cd66d227f43a3988060d4878dd7a8897

@MarkusFlorian79 I guess you have to make sure to use the 0.2.1 version of the image, either this, or you have some setup pulled from outside of the image.

Also the image version that I mention has this cmd:
"Cmd": [ "pipenv", "run", "gunicorn", "-b", "0.0.0.0:80", "httpbin:app", "-k", "gevent" ],

Thank you for accepting the PR @pmalek and thanks for putting the time and effort in maintaining an open source project. Respect!