Skip to content

Conversation

romanfurst
Copy link

Current implementation

since default Redis images are hardcoded, updating redis-operator version might cause all incances of redisdailover cluster are restarted if hardcoded default values has changed. It might be undesirable behavior especially in production environment (assume deploy with gitops argocd)



Propose implementation

default images are externally configurable. You can change redis-operator version but your defined default versions of redisfailover instances stay still -> redisfailover instances not change -> no restart

@romanfurst romanfurst requested a review from a team October 6, 2023 17:51
@romanfurst romanfurst force-pushed the default-images-config branch from aaefa34 to ee3db0a Compare October 6, 2023 19:13
@romanfurst romanfurst changed the title enable ovveride defaults for redis images enable override defaults for redis images Oct 6, 2023
@romanfurst romanfurst force-pushed the default-images-config branch from ee3db0a to f511aa8 Compare October 6, 2023 19:54
@SISheogorath
Copy link

SISheogorath commented Oct 16, 2023

Personal opinion: I would love to see a Full Qualified Image Name(FQIN, docker.io/library/redis:6.2.6-alpine instead of redis:6.2.6-alpine) since all main kubernetes runtimes (containerd, cri-o, …) allow to override mirroring of registries already, there is no benefit of not using a FQIN.

Some container runtimes can be configured to refuse to start when the image name is not a FQIN, e.g. cri-o with strict configuration. Further it helps to make clear what redis image is actually intended to be used and not just some image that happens to have these tags causing unexpected issues due to different resolution of image names/tags.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants