-
Bug
-
Resolution: Fixed
-
Minor
-
None
-
kubernetes plugin v0.12 or current master
Jenkins 2.62
container step running in a _debian_ container
podTemplate(name: "mypod", label: "label", containers: [ containerTemplate(name: 'debian', image: 'debian', ttyEnabled: true, command: 'cat', ) ]) { node("label") { container('debian') { sh 'for i in $(seq 1 1000); do echo $i; sleep 0.3; done' } } }
leads to
[Pipeline] podTemplate
[Pipeline] {
[Pipeline] node
Running on horst-nwmsn-32h5h in /home/jenkins/workspace/full
[Pipeline] {
[Pipeline] container
[Pipeline] {
[Pipeline] sh
[full] Running shell script
+ seq 1 1000
+ echo 1
1
+ sleep 0.3
+ echo 2
2
+ sleep 0.3
[Pipeline] }
[Pipeline] // container
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // podTemplate
[Pipeline] End of Pipeline
ERROR: script returned exit code -1
Finished: FAILURE
Sometimes it fails a bit faster.
Might be related to the script being started with "nohup" now.
- duplicates
-
JENKINS-47791 Eliminate ProcessLiveness
-
- Resolved
-
- links to
[JENKINS-46651] container step "script returned exit code -1"
| Issue Type | Original: New Feature [ 2 ] | New: Bug [ 1 ] |
| Environment |
New:
kubernetes plugin v0.12 or current master Jenkins 2.62 container step running in a _debian_ container |
| Priority | Original: Minor [ 4 ] | New: Trivial [ 5 ] |
Seems that the debian:latest container does not have ps by default, and durable-task uses it to check for the process.
Should we document this?
Btw. workaround is to derive your own image from "debian" and install the procps package in it.
I have similar problem with exit -1 error but i do have procps package installed and its happens only with 10+ concurrent build of same project.
kubernetes plugin v1.0
Jenkins 2.37
Here is my pipeline:
tag = env.BUILD_TAG
podTemplate(label: tag,
volumes: [
hostPathVolume(mountPath: '/var/lib/docker', hostPath: '/tmp/'+tag)
],
containers: [
containerTemplate(name: 'jnlp', image: 'localreg:5000/jenkinsci/jnlp-slave', args: '${computer.jnlpmac} ${computer.name}'),
containerTemplate(name: 'java', image: 'localreg:5000/base/java', ttyEnabled: true, command: 'cat'),
containerTemplate(name: 'dind', image: 'localreg:5000/base-builders/dind:1.0.0', ttyEnabled: true, privileged: true, alwaysPullImage: true)
]) {
node(tag) {
try {
container('jnlp') {
stage('Preparing') {
checkout scm
setGitEnvironmentVariables()
setSimpleEnvironmentVariables()
}
}
container('java') {
stage('Build a Maven project') {
sh 'env'
}
}
container('dind') {
stage('Build Docker') {
sh "docker build . || echo real bad exit code !"
}
}
} catch (error) {
throw error
}
}
}
Sometimes i receive this error while 10+ concurrent builds are running:
/home/jenkins/workspace/docker_builds_master-EHQCE25ZSFSE4QALBNHGET3HINBLMWZKHOXT7335PWTFXXMWNJDQ [Pipeline] sh [docker_builds_master-EHQCE25ZSFSE4QALBNHGET3HINBLMWZKHOXT7335PWTFXXMWNJDQ] Running shell script + docker build . Sending build context to Docker daemon 74.24 kB Step 1 : FROM localreg:5000/base/ubuntu-14.04 latest: Pulling from base/ubuntu-14.04 96c6a1f3c3b0: Pulling fs layer e8c945afff4f: Pulling fs layer b46adc58e13f: Pulling fs layer e8c945afff4f: Verifying Checksum e8c945afff4f: Download complete b46adc58e13f: Download complete 96c6a1f3c3b0: Verifying Checksum 96c6a1f3c3b0: Download complete [Pipeline] } [Pipeline] // stage [Pipeline] } [Pipeline] // container [Pipeline] } [Pipeline] // node [Pipeline] } [Pipeline] // podTemplate [Pipeline] End of Pipeline ERROR: script returned exit code -1 Finished: FAILURE
Any suggestions what can it be ?
Max k
added a comment - - edited I have similar problem with exit -1 error but i do have procps package installed and its happens only with 10+ concurrent build of same project.
kubernetes plugin v1.0
Jenkins 2.37
Here is my pipeline:
tag = env.BUILD_TAG
podTemplate(label: tag,
volumes: [
hostPathVolume(mountPath: '/ var /lib/docker' , hostPath: '/tmp/' +tag)
],
containers: [
containerTemplate(name: 'jnlp' , image: 'localreg:5000/jenkinsci/jnlp-slave' , args: '${computer.jnlpmac} ${computer.name}' ),
containerTemplate(name: 'java' , image: 'localreg:5000/base/java' , ttyEnabled: true , command: 'cat' ),
containerTemplate(name: 'dind' , image: 'localreg:5000/base-builders/dind:1.0.0' , ttyEnabled: true , privileged: true , alwaysPullImage: true )
]) {
node(tag) {
try {
container( 'jnlp' ) {
stage( 'Preparing' ) {
checkout scm
setGitEnvironmentVariables()
setSimpleEnvironmentVariables()
}
}
container( 'java' ) {
stage( 'Build a Maven project' ) {
sh 'env'
}
}
container( 'dind' ) {
stage( 'Build Docker' ) {
sh "docker build . || echo real bad exit code !"
}
}
} catch (error) {
throw error
}
}
}
Sometimes i receive this error while 10+ concurrent builds are running:
/home/jenkins/workspace/docker_builds_master-EHQCE25ZSFSE4QALBNHGET3HINBLMWZKHOXT7335PWTFXXMWNJDQ
[Pipeline] sh
[docker_builds_master-EHQCE25ZSFSE4QALBNHGET3HINBLMWZKHOXT7335PWTFXXMWNJDQ] Running shell script
+ docker build .
Sending build context to Docker daemon 74.24 kB
Step 1 : FROM localreg:5000/base/ubuntu-14.04
latest: Pulling from base/ubuntu-14.04
96c6a1f3c3b0: Pulling fs layer
e8c945afff4f: Pulling fs layer
b46adc58e13f: Pulling fs layer
e8c945afff4f: Verifying Checksum
e8c945afff4f: Download complete
b46adc58e13f: Download complete
96c6a1f3c3b0: Verifying Checksum
96c6a1f3c3b0: Download complete
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // container
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // podTemplate
[Pipeline] End of Pipeline
ERROR: script returned exit code -1
Finished: FAILURE
Any suggestions what can it be ? omegam: Can you try setting org.jenkinsci.plugins.durabletask.BourneShellScript.LAUNCH_FAILURE_TIMEOUT to something higher than the default 15 seconds, and see if that helps?
Note:
Example works with busybox container.