Restart pod on Podman when related information is changed in Devfile

[rm3l]

/kind user-story

User Story

As a user running odo against Podman, I want changes to my Devfile to be detected and applied, So that I can continuously iterate on my project without having to restart odo dev manually.

Restarting odo dev manually every time I edit something in the Devfile can be annoying and time-consuming.

Currently, it detects any changes to the Devfile and warns users that odo dev needs to be restarted (see #6336). Even changes that would not affect the running Pod (like changes to the build or run commands) are ignored.

Acceptance Criteria

• [ ] It should recreate the Pod on Podman if there are changes in the Devfile that would modify the Pod spec (e.g., changes in the container component, ...)
• [ ] It should not recreate the Pod on Podman if there are changes in the Devfile that would not modify the Pod spec (e.g., changes to the run command only)
• [ ] Similarly to how odo dev works against a cluster, it should notify users of the current Pod state
• [ ] Similarly to how odo dev works against a cluster, external factors that might affect the running Pod should also be handled by recreating the Pod if needed (but that can be a separate issue)
• For now, it should just recreate the Pod on Podman if there any changes in the Devfile
• Update documentation: remove this point from the "Podman Support limitations" page

Links

• Related Epic (mandatory): [Epic] Use odo with Podman (without Kubernetes/OpenShift cluster) [stage 2] #6578
• Related Issue: odo dev needs to handle changes in devfile.yaml #5728

/kind user-story

[valaparthvi]

@rm3l To add more information on this.

[rm3l]

/unassign
/remove-triage needs-information

Description and acceptance criteria updated. We can discuss and review this in the grooming call.

[rm3l]

@valaparthvi I'm also removing this from the current Sprint (236) as it is not pointed yet, and so that it shows up in the Grooming view in the Project.

[valaparthvi]

Grooming Call [April 20 '23]:
Current scope of the issue:

1. For now, delete and recreate the pod no matter where the change occurs.
2. Volumes can be used to maintain persistent data.
3. Port forwarding needs to be restarted.

Look into other ACs after we decide on a new design to separate resources and commands.

[valaparthvi]

Should be fixed by refactoring done in #6771.