by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jan 31 2019 22:50
    batmat review_requested #3874
  • Jan 31 2019 21:29
    basil opened #3875
  • Jan 31 2019 18:18
    nisarg1499 opened #3874
  • Jan 31 2019 18:12
    nisarg1499 edited #3873
  • Jan 31 2019 18:04
    nisarg1499 edited #3873
  • Jan 31 2019 18:03
    nisarg1499 opened #3873
  • Jan 31 2019 15:53
    batmat synchronize #3865
  • Jan 31 2019 15:24
    jglick commented #3861
  • Jan 31 2019 15:18
    nisarg1499 closed #3863
  • Jan 31 2019 15:18
    nisarg1499 commented #3863
  • Jan 31 2019 15:13
    oleg-nenashev commented #3863
  • Jan 31 2019 15:11
    nisarg1499 reopened #3863
  • Jan 31 2019 15:11
    nisarg1499 closed #3863
  • Jan 31 2019 15:01
    nisarg1499 synchronize #3863
  • Jan 31 2019 14:54
    batmat synchronize #3865
  • Jan 31 2019 14:45
    batmat synchronize #3865
  • Jan 31 2019 14:16
    Wadeck closed #3805
  • Jan 31 2019 14:12
    pjanouse commented #3802
  • Jan 31 2019 14:12
    pjanouse commented #3802
  • Jan 31 2019 14:05
    pjanouse synchronize #3802
christopher hahn
@chahn1138_gitlab

I am seeing it again:

CMake Error: The source directory "/var/lib/jenkins/workspace/centos-build-/soft/src" does not exist.

I have convinced myself that this is the TRIGGER_TYPE, but did not know that it was not possible to run a Project without one.

@fusion27 Can you run your actions as Powershell scripts? (think of Jenkins as just the driver....if you can script it, then Jenkins can run that script for you)
Björn Magnusson
@bjornmagnusson
@fusion27 I can recommend using flyway for that, pretty seamless. https://flywaydb.org/
Pigeast
@Pigeast
Hello, everyone, may I ask, now I want to ask Jenkins to send a request and report the result of its build after the completion of the build? I used the plug-in Outbound WebHook notification, but there is a problem, the url it needs must be written dead, and I can't get the environment variable, so I cancel it. Is there anything else?
image.png
anyother plugin could do it ?
christopher hahn
@chahn1138_gitlab
FYI: My incapacity regarding the TRIGGER_TYPE was confirmed. :0)
Casey Wise
@fusion27
thank you @bjornmagnusson !!
joshua-alday
@joshua-alday
Hey folks, apologies for being impatient. I'm currently getting tons of OOM issues from jenkins-to-aks-pod maven builds. Has anyone else encountered this issue before?
Marky Jackson
@markyjackson-taulia
What does the pod log output look like? Also what does your yaml look like. That is where I would start. Then look at the node the pod is running on
joshua-alday
@joshua-alday
So I'm using AKS on azure (I hope that doesn't make a huge difference.) The log output varies, sometimes the oom issue is consistent, but other times it's completely different. But normally related to resources.
can I post log snippets here?
Marky Jackson
@markyjackson-taulia
ue pastebin
joshua-alday
@joshua-alday
Here's one of the run logs and probably the more common issue.
https://pastebin.com/yANdsv7v
That's the yaml that jenkins sends to AKS for pod creation.
the yaml was copied from kubernetes using the view/edit yaml option.
Marky Jackson
@markyjackson-taulia
This is suspect
"resources": {
"requests": {
"cpu": "1200m",
"memory": "8Gi"
On the node that pod is running on what does the describe node look like?
It is a best practice to always set a limit with request
and that should be determined but the size of your cluster and pods running on a given node in a given cluster
To me it seems like this pod is requesting more then is available for that node and possibly for the cluster as a whole
joshua-alday
@joshua-alday
The reason I don't set limits, which I know I should be in general, is because jdk1.8 supposedly doesn't work well with limits.
but for the most part, limits exacerbates the issue.
meaning I've seen more failures including limits.
Marky Jackson
@markyjackson-taulia
I would look at what is running on that node and see if this yaml request is in line with what the node has available
From the surface it appears it does not
joshua-alday
@joshua-alday
it is. otherwise I couldn't get the pod to be created.
and I honestly don't need 8 gigs.
I have it up to 8 gigs to see if giving the pod more memory would fix the issue, but of course it doesn't
Marky Jackson
@markyjackson-taulia
are you able to run top nodes?
kubectl top nodes
joshua-alday
@joshua-alday
yes
Marky Jackson
@markyjackson-taulia
what does it show?
I would also run kubectl top pods --all-namespaces
joshua-alday
@joshua-alday
aks-agentpool-19896651-0 160m 4% 2133Mi 15%
that's both top nodes and pods.
Marky Jackson
@markyjackson-taulia
"name": "JAVA_TOOL_OPTIONS",
"value": "-XX:+UnlockExperimentalVMOptions -XX:+UseCGroupMemoryLimitForHeap -XX:+UseParallelGC -Xmx128m"
Possible ^
joshua-alday
@joshua-alday
Should I not use the VMOptions?
I don't believe it'll work with cpu shares without it.
So funny thing is. If I connect directly to the node and run the maven build, it works perfectly fine.
it seems that once the slave.jar is being utilized, something happens.
joshua-alday
@joshua-alday
but when I examine the process through JMX, everything looks fine.
Marky Jackson
@markyjackson-taulia
This is one of those things that you will have to dig down in the logs for. I would look at the kubelet logs also
joshua-alday
@joshua-alday
the kubelet logs only show me the the connection details really. Passed the ssh connection, I get no more logging.
Marky Jackson
@markyjackson-taulia
Did you try mocking this up in minikube to see if the same results are present?