Gliffy Diagram
Edit in Word
Plugin Information
Add VMware Lab Manager support to Jenkins DescriptionThis plugin add to Jenkins CI a way to control Virtual Machines hosted on VMware Lab Manager. You can configure a Jenkins Slave, selecting a virtual machine from a Organization / Workspace / Configuration triplet, in this way, when you need to build a Job on a specific Slave, this VM will be startup up and shutdown or suspended again after the build process. RequirementsTo rebuild the WSDL2Java created files you will need to install Axis2 and use the wsdl2java program. All runtime requirements are included. ConfigurationLab Manager cloudThe first step is to configure Hudson to know what configuration in Lab Manager you will be using. To do this you need to add a new "Cloud" in the Hudson "Configure System" menu. The required parameters to setup are:
If you need some other particular settings, you can click on the Advanced button to change the default parameters.
To verify all you parameters you can click on Test button and check the output reported. SlavesNow you can setup your nodes in hudson and use them to build your projects. Going ahead with configuration you can see a page that looks like the normal node creation page, with three combo box added. The first one where you have to select the Lab Manager instance (the brief description provided in the configuration section). In the second one you pick the name of the Virtual Machine in Lab Manager configuration that you are using. In the third drop down you select the action to be taken when the VM is idle (it is recommended to pick shutdown over suspend due to overhead in Lab Manager). If you then select the option to have the slave be taken on or offline based on demand. Note that if you select Shutdown or Shutdown and Revert for an idle behavior the slave will not be available immediately. It will however come online once demand is polled a second time. Next, if using a JNLP slave (ie for Windows), you must check the Force VM launch option. Doing this along with the normal best practices to have a Windows slave turn on and auto-start the JNLP client is all that is required specially for Windows. Finally, you can set the delay between telling Lab Manger to get the VM online and Jenkins attempting to connect to it as a slave. This value is in seconds. All slaves must have the VMware Tools installed in order for it to be cleanly managed. Change LogVersion 0.2.8 (Aug 05, 2011)
Version 0.2.7 (Jul 18, 2011)
Version 0.2.5 (Jul 8, 2011)
Version 0.2.4 (Jul 7, 2011)
Version 0.2.2, 0.2.3 (Jan 24, 2011)
Version 0.2.1 (Jan 22, 2011)
Version 0.2.0 (Dec 14, 2010)
Version 0.1.9 (Dec 13,2010)
Version 0.1.8 (Dec 08, 2010)
Version 0.1.7 (Dec 07, 2010)
|
Lab Manager Plugin
Skip to end of metadata
Go to start of metadata
Comments (3)
Jan 18, 2011
casetaintor - says:
I'm curious what configuration you used to connect to the slave computer. I hav...I'm curious what configuration you used to connect to the slave computer. I have it working so that it starts up and shuts down VMs in Lab Manager, but I'm unable to get it to actually start the Hudson slave process on these machines. Do you start it via SSH? Have you tried this on Windows? JNLP? Any help would be greatly appreciated.
Jul 19, 2011
mbaker000 - says:
From what I can gather, if it's a Windows slave you'll need to follow the instru...From what I can gather, if it's a Windows slave you'll need to follow the instructions described on this page. You have a couple options:
Unfortunately it doesn't look like it's so streamlined as to allow Jenkins nodes to be created and destroyed automagically (and allow driving by Labels only), so each Lab Manager node will need to be added to Jenkins manually and connect/disconnect as the same node each time. However, with some Jenkins CLI magic you might be able to accomplish this when the slave goes to start up.
Hope this helps!
Oct 07, 2011
Trevor Baker says:
It isn't clear to me how/if the plugin is deploying the vm. The cloud part...It isn't clear to me how/if the plugin is deploying the vm. The cloud part for getting the instance and vm name is working, but I seem to have to deploy the vm via lab manager manually to get the ip address, which I then need to use in the secondary launch ssh configuration. I had hoped that the plugin would log into lab manager, start a vm given the specified paramaters, and return a variable containing the IP that I could use in the secondary launch.
My company has strict lease policies on vms, and I have hoped to skirt that via this plugin's ability to start/stop depoy/undeploy. Can you clarify, and thanks for your efforts!
Add Comment