Kubernetes plugin - ready probe

[Romain Grécourt]

Hi,

I'm doing a Jenkins pipeline with the kubernetes plugin (1.12.4).

Currently facing some odd behaviors where steps are executed before the pod is actually "ready". See some of the errors  I got below.

One of the recurring error is about a non existing directory inside the workspace, or the main container not existing in the pod.

BTW, what's exit code -2 ?

Looking at the pod definition I don't see any probe for the injected jnlp container.

Should I define a probe to ensure a proper readiness ? Does anyone have recommendations on how to do this the right way ?

See my Jenkinsfile here https://github.com/romain-grecourt/ee4j-glassfish/blob/0cb4600c17f3a875c451b343444b9d57b3ff3283/Jenkinsfile ; any feedback is welcome.

Thanks,

Romain

----

[ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ] Running shell script

process apparently never started in /home/jenkins/workspace/ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ@tmp/durable-3fb9a8c3

script returned exit code -2

----

[ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ] Running shell script

container [glassfish-ci] does not exist in pod [jenkins-slave-p3p21-whthx]

----
[ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ] Running shell script

process apparently never started in /home/jenkins/workspace/ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ@tmp/durable-fecb2d64

sh: 1: cannot create /home/jenkins/workspace/ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ@tmp/durable-fecb2d64/jenkins-log.txt: Directory nonexistent

sh: 1: cannot create /home/jenkins/workspace/ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ@tmp/durable-fecb2d64/jenkins-result.txt.tmp: Directory nonexistent

mv: cannot stat '/home/jenkins/workspace/ee4j-glassfish_ci-pipeline-YWJPM77PBRNTKWL4JGYGVAYHJW5MFPG52VIUHMUAGFIJ2G7NYJHQ@tmp/durable-fecb2d64/jenkins-result.txt.tmp': No such file or directory

script returned exit code -2
----

[Carlos Sanchez]

you need to check your pod logs and k8s describe them to see what's the possible error

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/CADhPohneMZ5%3Da2GPb5chSsjjRYctaJaDQbEXuKhBALO-N83o_A%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Romain Grécourt]

Unfortunately I don't have cluster access (eclipse policy) and this issue is somewhat intermittent.

Will get back to you if/when I get these logs.

To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/CALHFn6PSQPDZ1h8n%2BKQ21EMqUnHKQmxrCn2JFHvjQurG032sFQ%40mail.gmail.com.

[Carlos Sanchez]

the latest agent logs may be in the jenkins logs if the agent fails to start

To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/CADhPohk06Ri4U4jYjARxTxKE-iE19sCWqLDC9p-Ws3-D2WX8LQ%40mail.gmail.com.