search

kinvolk / lokomotive

🪦 DISCONTINUED Further Lokomotive development has been discontinued. Lokomotive is a 100% open-source, easy to use and secure Kubernetes distribution from the volks at Kinvolk
https://kinvolk.io/lokomotive-kubernetes/
Apache License 2.0
320 stars 49 forks source link

Add --dry-run flag #337

Open rata opened 4 years ago

rata commented 4 years ago

When hacking lokomotive, it is really useful to have a "dry-run" option. Several times you just want to test something, without actually creating the infrastructure.

I end up adding a return nil here or there, to make it not create the infra when hacking locally. IMHO, it makes sense to have a --dry-run option to cluster apply.

invidian commented 4 years ago

What was mentioned some time ago was a render operation, which would create all the assets, which then can be picked up by the apply command. This way, user should be able to change all needed assets without recompiling lokoctl etc. As this issue seems similar to this idea, perhaps we should align both.

invidian commented 4 years ago

There is an issue for it actually: #189.

rata commented 4 years ago

I'm not sure that will cover this, though. For things like manual provider in packet you run terraform several times, and you may want to see what is run or done at different stages.

I agree that is useful too, but is not really the same, IIUC.

invidian commented 4 years ago

What would cluster apply with --dry-run do then? Could you please clarify?

rata commented 4 years ago

Yes :)