There are 2 possible ways (that I can think of):

   1. Use a set of scripts that runs on an agent. These scripts will
   connects to a VM provider (vmware, virtualbox, ...) to spin up VMs. Then
   connect to the created VM to install, configure and test whatever you need
   to test.
   2. Write an elastic agent plugin for your VM provider
   <https://plugin-api.gocd.org/current/elastic-agents/#elastic-agents> (or
   use an existing one <https://www.gocd.org/plugins/#elastic-agents>).
   Elastic agents can bring up agents in response to build workloads.

I'd think option 1 is probably what you may want to start with, before
diving into option 2.

- Ketan



On Thu, Aug 1, 2019 at 5:57 AM Denis Troller <denis.trol...@gmail.com>
wrote:

> We are using GO-CD to build and test our product under Windows. The
> product is a desktop application.
>
> We have a pipeline that builds the product and we use virtual machines
> (under HyperV) to test it in controlled environments. Each VM hosts a GOCD
> agent to carry out the installation and testing of our product.
>
> The system needs to reset the VMs by applying a snapshot before actual
> deployment and testing can occur.
> It is easy to restore the VMs using powershell.
>
> The big-picture pipelines are:
> - Build
> - Hardware-Tests (on a real machine with hardware attached): Copy files,
> Test Application
> - Software-Tests (on a VM): Restore VM, install Application, test
> - Customer-Specific tests (on a VM): Restore VM, install Application, test
>
> This set of pipelines is replicated for each maintenance branch of the
> product (trunk, v1.0.*, v1.1.*... v2.0.*)
>
> Ideally I would like to be able to have any number of VMs available for
> software and customer tests to allow concurrent builds of different
> versions of the product.
> My problem is in modelling this in GOCD since I do no want two pipelines
> to restore the same VM. GOCD can choose a VM agent to cary out the testing,
> but the agent cannot restore its own VM (since it will then disappear and
> be reset, leading to a pipeline failure).
>
> Any idea on how to deal with such an environment ? Or am I doing this
> wrong ? Has anyone implemented something like this before ?
>
> --
> You received this message because you are subscribed to the Google Groups
> "go-cd" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to go-cd+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/go-cd/6c3a3cd3-cfb6-469f-ab4a-2ed35b011e1d%40googlegroups.com
> <https://groups.google.com/d/msgid/go-cd/6c3a3cd3-cfb6-469f-ab4a-2ed35b011e1d%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>

-- 
You received this message because you are subscribed to the Google Groups 
"go-cd" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to go-cd+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/go-cd/CAMUPJd6S6t1Pjik3V8eyhHuFnZiRZERRBOH3tL1g2FjShU-Ung%40mail.gmail.com.

Reply via email to