New DSC v3 implementation for cloud + on premises

[alv000h]

Currently I need a tool for automation and Configuration Management i want make a POC of DSC v3.

Main premises are avoid cloud locking and Linux / Windows compatibility (cloud and on premise)

I am reviewing the DSC documentation and I am getting very confused because of mixing documents for v2 and v3 and usually not clarifications (eg. disclaimer pointing this is out-dated) and lacking of documentation for v3 specially for linux.

I feel that a quickstart guide covering all use case scenarios for DSC v3 will be very helpful, this should cover at least.

• Quick config for linux (dependencies, ports, configurations required, etc), making clear that nxtools is a "must"
• Quick config for windows (dependencies, ports, configurations required, etc)
• Real working examples for linux / windows (how execute bash / ps1 script will be a good example)
• Quick discussion about working modes: pull vs push
• How can implement a server for polling in pull mode, there is not documentation or present one have errors (recommending azure machine config when DSC v3 is currently unsupported or in preview)`

the main question is, after investigate several days and reading documentation and repositories, i cannot response the questions related to main working modes (pull vs push) and what is the solution to implement a pull server to work in pull mode.

Any answers related to this topic will be welcome... as commented in other thread I saw, I see little info oriented to non cloud locking community... this biased views make DSC v3 mostly unusable, because I would not to migrate all to azure or buy an azure subscription only to manage my private cloud that in other part could be non internet connected because of security. fire and motion, you know

[michaeltlombardi]

the main question is, after investigate several days and reading documentation and repositories, i cannot response the questions related to main working modes (pull vs push) and what is the solution to implement a pull server to work in pull mode.

I think the reason you're struggling to find information about this in our documentation is that unlike PSDSC, DSC doesn't currently include an option to run as an agent service. You always invoke DSC with the resources available on the machine and given a specific configuration document (or invoking resources individually with the dsc resource commands). I can certainly clarify this in the docs.

Regarding your other suggestions for documentation, and speaking only to my own experience/context:

• I do eventually want to provide more functional quickstarts for configuring Windows, Linux, and macOS machines, but have held off on those to prioritize other documentation while resources are implemented. Right now we can't really show much in the way of configuration without relying on PSDSC resources, which invites pointed questions and optics for Linux and macOS users.
• I am currently specifically avoiding authoring a document showing a user how to use DSC to execute script files as not-resources, because it runs counter to the purpose of idempotent configuration management.