Replies: 4 comments 10 replies
-
one module could be a salt fabric, and another could be vmware vcenter for example. might be able to keep all the salt or vmware specific code out of the core codebase. |
Beta Was this translation helpful? Give feedback.
-
This makes sense. We'd have to think through the design a bit and also think through the |
Beta Was this translation helpful? Give feedback.
-
I agree it would be helpful to have a plugin system becuase I REALLY don't want to limit the ways that parts can be deployed. From our earlier work on QHub Cloud it is clear to me that infrastructure and services that run on the infrastructure is a line that we need to draw to enable these pluggins. For example: In order for QHub Cloud to be cloud agnostic we have to provision some infrastructure:
The rest are the services deployed on the kubernetes cluster For QHub OnPrem there are things we assume about the infrastructure at the moment:
Similar to QHub Cloud all the services are deployed atop these machines. The nice advantage of QHub OnPrem is that the worker nodes for the cluster install minimal software/services so they can usually be tied to existing infrastructure at a company to bring them into the fleet. |
Beta Was this translation helpful? Give feedback.
-
I'm closing old discussions in a repo cleanup effort. Feel free to re-open if needed. |
Beta Was this translation helpful? Give feedback.
-
I propose adding in a “python plugin system to setup/deploy/config hosts for qhub-hpc”. I want to be able to run something like qhub “init gcp -foundation [some plugin]”. That will use a standard python interface to call a 3rd party module. That module does whatever it does to create/deploy hosts and it returns a list of ip addresses that qhub can ssh into. Then the module returns and qhub-hpc continues. I think we can make and support 1 mock/sample module and the community can make others. @dharhas
Beta Was this translation helpful? Give feedback.
All reactions