Reading DSC Future Direction Update - PowerShell Team I understand (guess) that:
- DSC Pull Server will be deprecated soon; there will be support but no new versions
- MS will encourage customers to use Automation DSC on Azure, even for on prem hosts
- If it’s not successfull, MS would propose an on prem version of Automation DSC
What do you think ?
I started working on a reporting solution for DSC Pull Server (open-source but not published yet ;-)), I don’t know what to do with it: continue because DSC on premise will remain, for a time, the realm of Pull Server (and if well designed, the solution could work with both Pull Server and Automation), or stop now because no one will need such a solution if MS publish a ‘real’ (finished) product …
I think that there will still be use cases for on-prem solutions. Primarily for those who are still adverse to going to the cloud for any infrastructure. So your solution would still be valid in my opinion for a time.
Also, have you considered contributing to Tug? GitHub - PowerShellOrg/tug: Open-source, cross-platform Pull/Reporting Server for DSC
Microsoft’s commitment to cloud solutions is first and foremost, so I don’t readily see an on-prem version of this becoming available. While there is Azure Stack, the solution is more designed for managing on prem workloads that a company may not be allowed to run in the cloud. Command and control, however, is something they’re pushing heavily to go out of the datacenter.
But who knows? They could change their mind. Things are very fluid nowadays. Which is exciting!
Keep in mind the original pull server was “sample code.” I’m not sure MS ever intended to support it in production long-term. I also question, just a teeny bit, on whether a “push server” (more a la SCCM) wouldn’t eventually be a better idea. But as Will points out, Tug is an open-source pull server that you can actually code up yourself to behave exactly as you want.
Thank you for your answers.
Yes, I consider contributing to TUG to enable the same reporting as on Pull Server. The solution I propose will not depend on who provides the data.
As soon as I have something to show I’ll comme back here.
Also, as I seen in another thread, extending TUG to generate MOF on demand is something that interests me.