我的Go服务器无法在Cloud Foundry中启动

Question

I have a simple GO server that I am trying to push to cloud foundry on Bosh lite ( Vagrant + Virtual box.)

Here is the source code setup from my go application:

• ~/workspace/src/github.com/me/ ( parent directory)

  -Godeps

  -weight weight.go <-- main file.

  manifest.yml

  Procfile

My weight.go is a simple server that listens on 9000.

1) manifest.yml looks like this.

applications:
- name: weight
  memory: 128MB
  instances: 1

2 ) Procfile looks like this.

worker: bin/weight

3 ) I use the default buildpack.

4 ) When I push my app with cf push weight -c "./bin/weight;sleep 1d" I get:

a-424e-b509-6df11fb32cc7 ({"state"=>"STOPPED"})
2016-03-13T11:02:45.70-0700 [DEA/0]      OUT Got staging request for app with id 16d3795a-8cda-424e-b509-6df11fb32cc7
2016-03-13T11:02:46.84-0700 [API/0]      OUT Updated app with guid 16d3795a-8cda-424e-b509-6df11fb32cc7 ({"state"=>"STARTED"})
2016-03-13T11:02:46.89-0700 [STG/0]      OUT -----> Downloaded app package (12K)
2016-03-13T11:02:48.76-0700 [STG/0]      OUT -----> Downloaded app buildpack cache (78M)
2016-03-13T11:02:48.82-0700 [STG/0]      ERR Cloning into '/tmp/buildpacks/go-buildpack'...
2016-03-13T11:03:05.66-0700 [STG/0]      OUT Submodule 'compile-extensions' (https://github.com/cloudfoundry/compile-extensions.git) registered for path 'compile-extensions'
2016-03-13T11:03:05.68-0700 [STG/0]      ERR Cloning into 'compile-extensions'...
2016-03-13T11:03:07.59-0700 [STG/0]      OUT Submodule path 'compile-extensions': checked out '26a578c06a62c763205833561fec1c5c6d34deb6'
2016-03-13T11:03:07.61-0700 [STG/0]      OUT -------> Buildpack version 1.7.3
2016-03-13T11:03:09.81-0700 [STG/0]      OUT https://pivotal-buildpacks.s3.amazonaws.com/concourse-binaries/godep/godep-v55-linux-x64.tgz
2016-03-13T11:03:09.88-0700 [STG/0]      OUT -----> Checking Godeps/Godeps.json file.
2016-03-13T11:03:09.92-0700 [STG/0]      OUT -----> Using go1.5.3
2016-03-13T11:03:09.92-0700 [STG/0]      OUT -----> Running: godep go install -tags cloudfoundry .
2016-03-13T11:03:11.19-0700 [STG/0]      OUT -----> Uploading droplet (2.0M)
2016-03-13T11:03:25.11-0700 [DEA/0]      OUT Starting app instance (index 0) with guid 16d3795a-8cda-424e-b509-6df11fb32cc7
2016-03-13T11:03:34.93-0700 [DEA/0]      OUT Removing crash for app with id 16d3795a-8cda-424e-b509-6df11fb32cc7
2016-03-13T11:03:34.93-0700 [DEA/0]      OUT Stopping app instance (index 0) with guid 16d3795a-8cda-424e-b509-6df11fb32cc7
2016-03-13T11:03:34.93-0700 [DEA/0]      OUT Stopped app instance (index 0) with guid 16d3795a-8cda-424e-b509-6df11fb32cc7
2016-03-13T11:03:56.81-0700 [DEA/0]      OUT Starting app instance (index 0) with guid 16d3795a-8cda-424e-b509-6df11fb32cc7
**2016-03-13T11:03:57.79-0700 [DEA/0]      ERR Instance (index 0) failed to start accepting connections**
2016-03-13T11:03:57.80-0700 [API/0]      OUT App instance exited with guid 16d3795a-8cda-424e-b509-6df11fb32cc7 payload: {"cc_partition"=>"default", "droplet"=>"16d3795a-8cda-424e-b509-6df11fb32cc7", "version"=>"2eeebbf8-e84e-412a-aeba-2adc0cffea6b", "instance"=>"910eba711c6e414bb7c6324565d0a9af", "index"=>0, "reason"=>"CRASHED", "exit_status"=>127, "exit_description"=>"failed to accept connections within health check timeout", "crash_timestamp"=>1457892237}

I have tried setting

cf set-env weight PORT 9000

No success, Am I doing something funky with the manifest and yaml ?

The weight.go runs locally.

Code :

func main (){
    http.HandleFunc("/weight", weightHandler)
    err:=http.ListenAndServe("localhost:"+getPort(), nil)
    if err != nil {
        fmt.Println("got an err ")
        log.Fatalln(err)
    }else{
        fmt.Println("Apparently it works ?")
    }
    //fmt.Println("Hi")
    //time.Sleep(1*time.Hour)
}

func getPort() string {
    var port string
    if port = os.Getenv("PORT"); len(port) == 0 {
        fmt.Println("Didn't Found it")
        fmt.Println(port)
        port = DEFAULT_PORT
    }else{
        fmt.Println("Gotim")
        fmt.Println(port)
    }
    return port
}

I do get the Log for the random PORT CF assigns, still stuck at

2016-03-13T16:58:40.90-0700 [API/0]      OUT App instance exited with guid e5d417bd-c38d-4239-aa61-e9ca67fce79a payload: {"cc_partition"=>"default", "droplet"=>"e5d417bd-c38d-4239-aa61-e9ca67fce79a", "version"=>"8d7e80b3-69d5-4c83-9d37-1159d5deeba8", "instance"=>"14223d325c204406b87a131c065c16cc", "index"=>0, "reason"=>"CRASHED", "exit_status"=>-1, "exit_description"=>"failed to accept connections within health check timeout", "crash_timestamp"=>1457913520}

*FIXED IT : The problem was listening on localhost, I didn't specify the IP and I was good to go *

Answer 1

From Cloud Foundry docs on troubleshooting:

Make sure your application code uses the PORT environment variable. Your application may be failing because it is listening on the wrong port. Instead of hard coding the port on which your application listens, use the PORT environment variable. [Source]

From Cloud Foundry docs on environment variables:

The port on which the application should listen for requests. The Cloud Foundry runtime allocates a port dynamically for each instance of the application, so code that obtains or uses the application port should refer to it via the PORT environment variable. [Source]

So rather than your app being hardcoded to listen on port 9000, it should listen on the port specified by the PORT environment variable, and can then default to 9000 if that environment variable is not set.