Replies: 6 comments 8 replies
-
Idea 1 (obvious): just let patch versions diverge. Ironic-image already does not match Ironic versions perfectly. |
Beta Was this translation helpful? Give feedback.
-
Idea 2: use a sub-patch version (28.0.1.1 etc). Are we going to annoy people with that? |
Beta Was this translation helpful? Give feedback.
-
Idea 3 (also obvious): use independent versioning. How do people know which version of the operator installs which Ironic? |
Beta Was this translation helpful? Give feedback.
-
For example postgres-operator supports a range of postgres versions, it is not installing a single major version of postgres, much less matching versioning to minor or patch level. Since the operator is a lot of code that is not related to the product it operates, I'm voting for independent versioning. |
Beta Was this translation helpful? Give feedback.
-
not a lot to add here what Tuomo already wrote, I'm also for independent versioning as it just makes sense |
Beta Was this translation helpful? Give feedback.
-
@tuminoid @elfosardo okay, noted. I'm curious how supporting several versions may work in combination with custom images. If an operator or a downstream is using non-upstream image, should we: (a) assume version==latest, (b) expect them to provide a version mapping somehow (new CRD?). |
Beta Was this translation helpful? Give feedback.
-
The initial idea (I think I put that to the design proposal) was to couple IrSO versions with ironic-image versions. This makes sense from the good practices perspective (an operator installs a specific version of software and handles upgrades). But also, if we need to make point releases of IrSO specifically, the versions may start diverging.
Beta Was this translation helpful? Give feedback.
All reactions