feature-request: Delayed destruction

[akutz]

Hi All,

Due to some previous issues, as well as the current #161, we're trying to figure out the best method to delay the destruction of our environments. PR #154 provides the ability to destroy an environment without the need for the Terraform state, so one method would be to simply disable the destroy step for vSphere and let us manually remove resources a day or two later.

However, I think this is a feature that may also add value to other platforms. The fact is, it's hard to debug failed deployments when the resources are no longer present. Logs may be attached to those resources, and when the resources are destroyed, so too are the logs.

This is potentially a priority issue for VMware.

Thank you!

[akutz]

Hi All,

After a few hours round-the-bin with internal support here at VMware, they've concluded there is little they are able to do unless there are logs associated with failed deployments. I am going to submit a PR to make the destruction of vSphere environments dependent upon an environment variable. The default value will be to not destroy the environment.

[taylor]

Reference pull request for skipping destroy #163

[taylor]

Please change the default to be destroy.

If an environment variable VSPHERE_DESTROY_SKIP is set to true then destroy will be skipped.

Ideally this can be applied to all providers. Then we can turn on skipping cleanup when we want to debug.

Summary:

• Update code to skip destory if VSPHERE_DESTROY_SKIP == true
• Note: Default is destroy and environment var must be set

[akutz]

Hi @taylor,

Thank you for your feedback. Done and done. I also made it so DESTROY_SKIP is checked in case you want to generalize the feature.