setting multiple puppet environments within one VMware cluster

[maxadamo]

Hi,

first of all, I am not quite sure if this is the right place to raise an issue with this topic.

We are setting up cloudbolt in our infrastructure and we got to know from you, that puppet environments are necessarily linked to VMware clusters.
This behavior is seen as a limitation in most real life situations (and I make this consideration based on my experience with Puppet in several organizations).
We may agree that it's better to split the production environment, but you should not force you customers to squash all their environment, because they don't have enough clusters.
This same problem arises with the test, uat, devel environments. I have never seen so many clusters per puppet environments, and it would really lack the necessary flexibility.

Possible solutions:

• linking environments to VMware folders (instead of clusters)

• behave like Foreman does (this is in my opinion the best option). With Foreman you can store the environment that the VM sends when puppet runs for the first time. Foreman will always re-use that environment (regardless of cluster and folder and avoiding that the environment can be changed on the client side). I guess that Foreman stores these information on Postgres, and you can do the same with the backend used with Cloudbolt.

[LAKostis]

Citing foreman documentation about (I believe you're talking about this option):

enc_environment

When this is true, Foreman will send the puppet environment in the ENC yaml output. This is meant to fix conflicts between a node’s puppet.conf environment and the environment set in Foreman. On Puppet 3+, agents will take the environment sent by the ENC. When false, the ENC yaml will not contain the environment, the node will not update its environment and use the one at puppet.conf. Default: true