Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-41316

docker.image('my-image').inside{...} no longer honors Dockerfile "entrypoint" since version 1.8

    XMLWordPrintable

Details

    • Bug
    • Status: Reopened (View Workflow)
    • Major
    • Resolution: Unresolved
    • docker-workflow-plugin
    • None
    • docker-pipeline-plugin:1.9.1
      Jenkins 2.19.4.2

    Description

      The docker pipeline plugin used to honor the Dockerfile "entrypoint" until v1.8 in "docker.image('my-image').inside{...}". Since then, "entrypoint" is ignored.

      The capability to do selenium tests with a selenium-standalone-server running inside the Docker container fails since 1.8.

      This regression seems to be caused by JENKINS-37987 and the github commit "[FIXED JENKINS-37987] Override ENTRYPOINT, not just command, for WithContainerStep".

      This Jira seem to be similar to JENKINS-39748

      Testcase

      Code

      node ("docker") {
          docker.image('cloudbees/java-build-tools:2.0.0').inside {
      
              // verify that selenium-standalone-server has been started by the Dockerfile entrypoint /opt/bin/entry_point.sh
              sh "curl http://127.0.0.1:4444/wd/hub"
              
              // test with selenium python
              writeFile (
                  file: 'selenium_remote_web_driver_test.python', 
                  text: 
      """#!/usr/bin/env python 
      
      from selenium import webdriver
      from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
      
      driver = webdriver.Remote(
         command_executor='http://127.0.0.1:4444/wd/hub',
         desired_capabilities=DesiredCapabilities.FIREFOX)
      
      driver.get('http://python.org')
      assert 'Python' in driver.title
      """)
      
              sh "python selenium_remote_web_driver_test.python"
      
          }
      }
      

      Console

      The check "curl http://127.0.0.1:4444/wd/hub" will fail.

      Started by user admin
      [Pipeline] node
      Running on agent-1 in /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2
      [Pipeline] {
      [Pipeline] sh
      [testSelenium2] Running shell script
      + docker inspect -f . cloudbees/java-build-tools:2.0.0
      .
      [Pipeline] withDockerContainer
      $ docker run -t -d -u 1000:1000 -w /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2 -v /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2:/home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2:rw -v /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2@tmp:/home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2@tmp:rw -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** --entrypoint cat cloudbees/java-build-tools:2.0.0
      [Pipeline] {
      [Pipeline] sh
      [testSelenium2] Running shell script
      + curl http://127.0.0.1:4444/wd/hub
        % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                       Dload  Upload   Total   Spent    Left  Speed
      
        0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0curl: (7) Failed to connect to 127.0.0.1 port 4444: Connection refused
      [Pipeline] }
      $ docker stop --time=1 c65380bfd6c83d2290fc2e8fa8e5ae4cb0b84d1b21c66b4a3019c4a831f8833c
      $ docker rm -f c65380bfd6c83d2290fc2e8fa8e5ae4cb0b84d1b21c66b4a3019c4a831f8833c
      [Pipeline] // withDockerContainer
      [Pipeline] }
      [Pipeline] // node
      [Pipeline] End of Pipeline
      ERROR: script returned exit code 7
      Finished: FAILURE
      

      Workaround

      The workaround is to manually start the Dockerfile entrypoint with sh "nohup /opt/bin/entry_point.sh &".

      Code

      node ("docker") {
          docker.image('cloudbees/java-build-tools:2.0.0').inside {
              // WORKAROUND: MANUALLY START THE DOCKERFILE ENTRYPOINT
              sh "nohup /opt/bin/entry_point.sh &"
              sh "sleep 5"
              
              // verify that selenium-standalone-server has been started by the Dockerfile entrypoint /opt/bin/entry_point.sh
              sh "curl http://127.0.0.1:4444/wd/hub"
              
              // test with selenium python
              writeFile (
                  file: 'selenium_remote_web_driver_test.python', 
                  text: 
      """#!/usr/bin/env python 
      
      from selenium import webdriver
      from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
      
      driver = webdriver.Remote(
         command_executor='http://127.0.0.1:4444/wd/hub',
         desired_capabilities=DesiredCapabilities.FIREFOX)
      
      driver.get('http://python.org')
      assert 'Python' in driver.title
      """)
      
              sh "python selenium_remote_web_driver_test.python"
      
          }
      }
      

      Console

      Started by user admin
      [Pipeline] node
      Running on agent-1 in /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2
      [Pipeline] {
      [Pipeline] sh
      [testSelenium2] Running shell script
      + docker inspect -f . cloudbees/java-build-tools:2.0.0
      .
      [Pipeline] withDockerContainer
      $ docker run -t -d -u 1000:1000 -w /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2 -v /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2:/home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2:rw -v /home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2@tmp:/home/ubuntu/jenkins-aws-home/workspace/tests/testSelenium2@tmp:rw -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** --entrypoint cat cloudbees/java-build-tools:2.0.0
      [Pipeline] {
      [Pipeline] sh
      [testSelenium2] Running shell script
      + nohup /opt/bin/entry_point.sh
      [Pipeline] sh
      [testSelenium2] Running shell script
      + sleep 5
      [Pipeline] sh
      [testSelenium2] Running shell script
      + curl http://127.0.0.1:4444/wd/hub
        % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                       Dload  Upload   Total   Spent    Left  Speed
      
        0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
        0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
        0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
      [Pipeline] writeFile
      [Pipeline] sh
      [testSelenium2] Running shell script
      + python selenium_remote_web_driver_test.python
      [Pipeline] }
      $ docker stop --time=1 804a1f9cac0e8040b5e882a7c3ebd052df53e9cb99b34c0a7ffba4d0abff5401
      $ docker rm -f 804a1f9cac0e8040b5e882a7c3ebd052df53e9cb99b34c0a7ffba4d0abff5401
      [Pipeline] // withDockerContainer
      [Pipeline] }
      [Pipeline] // node
      [Pipeline] End of Pipeline
      Finished: SUCCESS
      

      Attachments

        Issue Links

          Activity

            The 1.15 release really screwed up all of our builds involving containers. Reverted to 1.14.

            starefossen Hans Kristian Flaatten added a comment - The 1.15 release really screwed up all of our builds involving containers. Reverted to 1.14.
            nre_ableton Nik Reiman added a comment -

            Ditto here, 1.15 has caused a bunch of our container-related builds to fail. Looking at the diff of https://github.com/jenkinsci/docker-workflow-plugin/pull/116/files, it seems that the "detection" is looking for `cat` commands, which we are not using within `docker.image.inside`.

            Sadly the `–entrypoint ""` workaround also does not work in our case and we have reverted to 1.14 for now. Is there another JIRA issue already to track this regression, or should this one be re-opened?

            nre_ableton Nik Reiman added a comment - Ditto here, 1.15 has caused a bunch of our container-related builds to fail. Looking at the diff of https://github.com/jenkinsci/docker-workflow-plugin/pull/116/files,  it seems that the "detection" is looking for `cat` commands, which we are not using within `docker.image.inside`. Sadly the `–entrypoint ""` workaround also does not work in our case and we have reverted to 1.14 for now. Is there another JIRA issue already to track this regression, or should this one be re-opened?

            Please don't re-open this issue, regression has been caught already and addressed in JENKINS-49278

            ndeloof Nicolas De Loof added a comment - Please don't re-open this issue, regression has been caught already and addressed in  JENKINS-49278
            sachin771 sachin gupta added a comment -

            I'm still facing the same issue with Jenkins version 2.161,

            I have search many issues opened and closed without any proper resolution to this problem, please provide the suitable resolution

            sachin771 sachin gupta added a comment - I'm still facing the same issue with Jenkins version 2.161, I have search many issues opened and closed without any proper resolution to this problem, please provide the suitable resolution
            faucherb94 Ben Faucher added a comment -

            I just ran into this issue. Is this going to be addressed? Why is the entrypoint overridden in the first place?

            faucherb94 Ben Faucher added a comment - I just ran into this issue. Is this going to be addressed? Why is the entrypoint overridden in the first place?

            People

              Unassigned Unassigned
              cleclerc Cyrille Le Clerc
              Votes:
              8 Vote for this issue
              Watchers:
              32 Start watching this issue

              Dates

                Created:
                Updated: