Replies: 1 comment 1 reply
-
I too am having a similer problem with MetalLB |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi,
I have a cluster running Flux where I am deploying resources for which part of their configuration changes dynamically through other tools.
I would like to ignore those specific changes during reconciliation, but continue to manage the rest of the manifest's fields with Flux.
An example use-case I have for this is when using KubeVirt.
Some VMs'
running
field should be set by Flux at initial apply, but should then be ignored so that it does not shut down/start up at unexpected times.Furthermore, some resources may change at runtime. For example, new persistent volumes can be mounted during backups, and RAM/CPU can be increased during intensive tasks.
These changes are short-term and the "usual" configuration should still be defined in Flux to be applied at initial deployment, but should not be reverted if it were to change in-cluster.
Other unchanging fields (node selectors...) should also continue being managed by Flux.
According to the docs, I currently have the choice between:
kustomize.toolkit.fluxcd.io/reconcile: disabled
kustomize.toolkit.fluxcd.io/ssa: merge
But none of these solutions allow to deploy initial values AND continue to manage other fields at the same time.
Is my understanding correct? Or is there a way to achieve what I am looking for?
Otherwise, is it because this goes against some of the core principles of Flux?
If this is a proposal which could be considered, I would be glad to help with its implementation.
Beta Was this translation helpful? Give feedback.
All reactions