-
Bug
-
Resolution: Fixed
-
Blocker
-
None
-
Jenkins Version: 2.121.1
Docker Pipeline Plugin: 1.5
-
-
1.24
After Upgrading Docker Pipeline Plugin from 1.4 to latest(1.7), our pipeline job running with Docker started failing with following error,
ERROR: The container started but didn't run the expected command. Please double check your ENTRYPOINT does execute the command passed as docker run argument. See
https://docs.docker.com/engine/userguide/eng-image/dockerfile_best-practices/#entrypoint
for entrypoint best practices.[Pipeline]
$ docker stop --time=1 6c3e4c47264e7766c2277555ac05de7267ab0895ab8f91b257761a6c649379c2
$ docker rm -f 6c3e4c47264e7766c2277555ac05de7267ab0895ab8f91b257761a6c649379c2[Pipeline] // withDockerContainer[Pipeline] }[Pipeline] // stage[Pipeline] }[Pipeline] // withDockerRegistry[Pipeline] }[Pipeline] // withCredentials[Pipeline] }[Pipeline] // ws[Pipeline] }[Pipeline] // node[Pipeline] echoError: hudson.AbortException: script returned exit code -2
Investigating further we found that issue is with Docker Pipeline Plugin 1.5.
Can you please help us fix this issue? Are we missing something while using version 1.5 or is it a bug in the plugin?
- causes
-
-
- Closed
-
- is duplicated by
-
-
- Reopened
-
-
-
- Closed
-
[JENKINS-52264] Error while running pipeline job in docker
I had the same issue and it was indeed an empty environment set in the global configuration. There was code added to protect from that in version 1.12 but was done only for docker run and not for docker exec. It is a shame really because the fix is a one line change.
This commit has the fix for the docker run and the same can be made for the docker exec in the same class.
https://github.com/jenkinsci/docker-workflow-plugin/commit/a25fe7ad582eecb771292084acdbff5ee56cd2c2
I created https://github.com/jenkinsci/docker-workflow-plugin/pull/147 which addresses this issue the same way it was done for "docker run" in the original commit mentioned above. Since that original commit was done the "docker exec" was changed to use the "–env" option which exposed this defect.
Jonathan Chauncey
added a comment - I was having this same issue and found that if you go Manage Jenkins -> Configure System there will be an empty environment variable in the Global Properties section. If you remove that everything should work
Jonathan Chauncey
added a comment - I was having this same issue and found that if you go Manage Jenkins -> Configure System there will be an empty environment variable in the Global Properties section. If you remove that everything should work In my scripted pipeline -v /var/jenkins/caches:/var/jenkins/caches it works as work-around.
docker.image("....").inside("{{-v /var/jenkins/caches:/var/jenkins/caches")
{
...
}
Giorgio Desideri
added a comment - - edited In my scripted pipeline -v /var/jenkins/caches:/var/jenkins/caches it works as work-around.
docker.image("....").inside("{{-v /var/jenkins/caches:/var/jenkins/caches ")
{
...
} This can manifest in various ways - for me it first appeared to be this bug, and only after enabling LAUNCH_DIAGNOSTICS did we see evidence that lead us here, which was the actual culprit.
I found help here for a similar issue: http://jenkins-ci.361315.n4.nabble.com/withDockerContainer-in-pipeline-failing-td4932626.html
Basically, in case the link dies: you likely have an empty environment variable in your global configuration or node configuration.