-
Notifications
You must be signed in to change notification settings - Fork 76
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Allow removing Bundles whilst keeping the synced CA certs #99
Comments
Issues go stale after 90d of inactivity. |
/remove-lifecycle stale |
Issues go stale after 90d of inactivity. |
/remove-lifecycle stale |
At the moment, each target configmap has an owner reference to the
Bundle
for which it was created.This means that if a
Bundle
gets deleted, the configmaps will be deleted by Kubernetes garbage collector.This can be useful for folks who sync them to a large number of namespaces and want to be able to easily delete
Bundle
and all related configmaps.Judging from our experience with cert-manager, it is likely that in the future there will be scenarios where users end up uninstall and re-install trust manager including the Bundle CRD, either to perform a more complex upgrade or to perform some other cluster component configuration change like moving to a different deployment mechanism etc.
If configmaps with CA certs get deleted it means that
This would mean that it is not safe to re-install trust-manager as it can cause downtime.
I think we should make it possible to re-install trust-manager without deleting the synced CA certs whilst still giving users an easy way to delete Bundles and associated configmaps.
We could:
At the moment owner references are used heavily by trust controller (to filter configmaps to watch, list etc). If we are going to get rid of them or make them optional, we'd need to devise another way how to filter target configmaps. One way could be labelling. Server side apply could also help- we could look at managed fields to determine whether a configmap was added by trust manager #89 . Also, if we are going to do it, it'd be easier now than when there are more users and stability guarantees.
The next step for this issue would be to come up with a design.
The text was updated successfully, but these errors were encountered: