We have Jenkins installed and running on a WebSphere Application Server. We recently upgraded the server to version 8.5.5 and switched the profile to use JDK version 1.7. Doing this Jenkins Crashes the WebSphere Application Server and we cannot tell why. Any hints or suggestions on things to look at? Switching the server back to JDK 1.6 seems to work just fine, can Jenkins not run on JDK 1.7 or is it something else?
I'll assume that by crash the websphere application server you mean either a fatal error at startup or a java process crash. Those can have many causes.
Just to give some ideas, it may be related to the fact that you had somewhat customized your JDK install and forgot to re-apply those customization to your new JDK. Or that switching SDK requires you to switch command line options, or that you indeed hit an incompatible class in the stack, or that your process crash because of bad luck, etc.
So please find more information in the logs, either the corresponding stack traces in your WAS server logs or the javacores crash files.
Please also report your jenkins version.
As for JDK 7 compatibility, latest jenkins itself should be compatible, yet some plugins are not
You may also want to read this: https://stackoverflow.com/questions/17411717/jenkins-on-websphere-reports-java-lang-noclassdeffounderror-jenkins-model-jenki. Maybe you have the same issue.
If you indeed find out an incompatibility, please report an issue in jenkins issue tracker and consider updating the Jenkins Websphere wiki.
Related
I am trying to upgrade Jenkins from version 2.176.1 to 2.375.2. I tried two methods and both of them failed.
Method-1: Download the jenkins.war and copy it into /usr/lib/jenkins/ directory by replacing the original one. After copying, when I restart Jenkins, it simply fails to restart.
Method-2: Upgrade Jenkins' RPM using yum update jenkins. This command runs successfully and process exists without any error. But, again when I restart Jenkins, it fails to start.
CentOS version: CentOS-7
First of all - I can only support opinion of Dmitriy Tarasevich about reading upgrade guides. You can find theme here: https://www.jenkins.io/doc/upgrade-guide/
Then - stick to the version upgrade process that you've used previously (or for the first time with installation). If you did install it via .*war, don¨t worry and use your Method-1. On the other hand, if you used yum, stick to that method-2.
Then - such a "big" upgrade at once from 2.176.1 to 2.375.1 is, I'd say, the reason it's failing for you as the version in between contain several major/breaking changes (our team has faced the same problem (we "just" wanted to bump the versions too much after a long time of no upgrading and we've faced several problems we weren't able to solve together). So it was necessary to split it into several independent upgrades which appeared to be doable and we managed to get to LTS 2.375).
To summarize recommended steps - I'd you need to:
1/ Split your upgrade into several steps and upgrade LTS more gradually: try to upgrade to first to some not-current LTS version. I believe something less than or precisely 2.303.3 could work and see if you can start it.
then actually run Jenkins and check administration page,
upgrade your plugins (if you want). FYI: this is recommended officially in Jenkins docs after every Jenkins core upgrade.
check Jenkins administration page and apply necessary changes / manage necessary decision that might appear there (if applicable).
2/ Try to upgrade to a higher LTS version (repeat step 1 with its sub-steps)
To be a very specific and give some examples (although I don't remember everything by hearth):
you skipped a gradually upgrade to version 2.319 LTS that brings a major change of working with master node --> built-in node that needs to be applied (confirmed) from within the administration section of Jenkins (see docs: https://www.jenkins.io/doc/book/managing/built-in-node-migration/ )
version 2.332.1 LTS brings an important change. I would say that you've hit exactly this issue (covered here): https://community.jenkins.io/t/jenkins-cant-find-java-after-update-running-on-centos-7/3615 - exact steps and more resources are within that thread and also here: https://www.jenkins.io/doc/upgrade-guide/2.332/#upgrading-to-jenkins-lts-2-332-1
from version 2.361 LTS Jenkins requires Java 11 or higher (see docs https://www.jenkins.io/doc/upgrade-guide/2.361/ ). So - you will need to ensure that you have a proper Java installed (you already might have it though). Including your agents!. Otherwise you would start your Jenkins but it would not connect any agent (usually undesired state).
Please, try to make some gradual progress and let us know where you managed to get
I installed Jenkins previously but it had some issue (there was no start icon) in the folder, so I uninstalled it. Now I'm trying to install it again, but it tends to pop this error:
I tried all the solutions mentioned in this post , but no success - Jenkins failed to start - Verify that you have sufficient privileges to start system services
As mentioned in the option 2 of the post, I tried to start jenkins in services but got a 1067 error
I read a lot of similar posts, but none of them really helped. I believe my user has enough privileges to work without causing any trouble, so the underlying problem is something else. Please help me figure out the issue.
I found the solution!
Turns out Jenkins doesn't work on other JDKs except for 8,11. And I have the 14th version. I'm putting the link to the answer that helped me change the JDK version.
Jenkins requires Java versions [8, 11] but you are running with Java 16 from C:\Program Files\Java\jdk-16.0.1
I want to use Vagrant to ensure a consistent development environment for my Groovy and Grails projects. Is there are standard base Vagrant box? Ie that comes with Java, GVM, Groovy and Grails pre-installed?
I’ve taken a look and can find two but I had some issues with both:
http://borzacchiello.it/set-up-a-vagrant-machine-for-grails-development/
“An error occurred while downloading the remote file. The error message, if any, is reproduced below. Please fix this error and try again. Couldn't open vag2demo/vagrant-grails/base”
https://github.com/osoco/testbox
Here localhost:8888 went to a tomcat that should have Jenkins running, instead it was a base tomcat installation.
Does anyone know of any other sources for a Groovy and Grails Vagrant box?
I ran into similar problems recently and created my own Vagrant box for Grails development.
https://github.com/janmey/grails-vagrant
It's based on Java7 and the latest Groovy and Grails versions are installed using gvm. It's still work-in-progress, but you should be able to get started with it pretty quickly. The inital "vagrant up" will take a while, but the "Hello World" app should work then.
I have been trying for sometime to get the TFS plugin working, and have had "semi" success.
The parts of the process that use Team Explorer command line client work well (I have defined the TFSPROXY environment variable, which seems to work.)
As soon as the plugin gets to the piece that uses the SDK, I get a stack trace printed, with the main error:
com.microsoft.tfs.core.exceptions.TECoreException: Unknown host somehost.on.corporate.network.
I have tried using the environment variables: HTTP_PROXY, and TFSPROXY. I have also tried adding the registry keys at HKLM/Software/MS/VisualStudio/10.0/TeamFoundation/SourceControl/Proxy (also added v9.0).
My question is: Is there any way to configure proxy settings for the SDK that the plugin is using, or has someone else had success with another alternative?
For anyone else who may be having this issue, I found that it is only the latest versions of the plugin using the SDK (3.0+), and the previous versions (2.0 and less) are using only the command line tool.
So as a workaround, I have downgraded to 2.0 and everything seems to be working correctly now.
We have recently installed Orbeon Forms stable 3.8.0 CE's orbeon.war and that works out of box.
Because some features do not work in the stable version we installed the CE nightly build orbeon-CE.war, but this one does not work out of the box.
Form builder has some exceptions. Not only the builder fails but the examples too, so Form runner too.
Java exception
java.lang.NullPointerException
class: java.util.zip.Deflater
method: ensureOpen
line: 421
We are using Linux Debian Lenny 2.6.26-smp, Tomcat 5.5.
Probably we are running into some undocumented requirement for this nightly build...?
Problem identified and workaround:
We use IBM java 1.5 and this is what we found.
Orbeon stable works OK with IBM java 1.5, Orbeon nightly does not work with IBM java 1.5, see previously mentioned zip error.
We then installed SUN java 1.5 and Orbeon nightly works OK with SUN java 1.5., but first there was a problem:
*Exception in thread "exist_QuartzScheduler_QuartzSchedulerThread" java.lang.OutOfMemoryError: PermGen space*
And we googled up this thread:
http://orbeon-forms-ops-users.24843.n4.nabble.com/Data-lost-on-quot-Save-Document-quot-td40450.html
The permgen space is a separate VM
setting. It can be increased with a
VM option, e.g.:
-XX:MaxPermSize=128m
Which solved the problem.
This is not a known issue, and this is something I am unable to reproduce here. Maybe the download you got was corrupted. Try to unzip the war file by hand (e.g. with unzip), and if this gives you an error, re-download a new war. If this doesn't solve your problem, could you update your question with more information, in particular the full stack trace for that NPE?