All notable changes to this project will be documented in this file.
- Generate OLM bundle for Release 23.4.0 ([#271]).
- Added support for converting secrets (including generating PKCS#12 bundles) (#286).
operator-rs
0.27.1
->0.44.0
(#275, #294).- Removed dummy key from generated Kerberos keytab (#285).
- [BREAKING] Daemonset for SecretOperator now assign resource requests and limits to all containers and init containers. Users who have configured resource limits previously in the 'values.yaml' file will need to move the configured limits from
.resources
to.node.driver.resources
for them to be honored going forward (#289).
- Added
kerberosKeytab
provisioner backend using MIT Kerberos (#99, #257). - Added experimental unprivileged mode (#252).
- Shortened the registration socket path for Microk8s compatibility (#231).
- Made kubeletDir configurable (#232).
- Microk8s users will need to
--set kubeletDir=/var/snap/microk8s/common/var/lib/kubelet
.
- Microk8s users will need to
- operator-rs: 0.25.0 -> 0.27.1 (#212).
- Include chart name when installing with a custom release name (#153).
- operator-rs: 0.10.0 -> 0.25.0 (#180).
- "privileged" security context constraints for OpenShift clusters (#144)
- Pods that consume Node-scoped
k8sSearch
secrets will now only be scheduled to Nodes that have the secret provisioned (#125).- This is only supported for pods that use the new-style
ephemeral
volume definitions rather thancsi
.
- This is only supported for pods that use the new-style
- Pods that consume secrets should now use the
ephemeral
volume type rather thancsi
(#125).csi
volumes will keep working for now, but should be considered deprecated, and will not be compatible with all new features.
- Pods that use
autoTls
volumes are now evicted when their certificates are about to expire (#114, commons-#20).
autoTls
CA generation now requires opt-in (#77).- The default
tls
SecretClass
now has this opt-in by default.
- The default
k8sSearch
backend's optionsecretLabels
has been removed (#123).
This release will cause any Pods that already used it get stuck Terminating when they are next deleted. The easiest way to fix this is to perform a rolling reboot of all nodes after the upgrade.
This is a one-time migration.
- Store secrets on tmpfs (#37).
- Locked down secret permissions by default (#37).
- Operator-rs: 0.8.0 -> 0.10.0 (#49).
- Fixed thread starvation and slow shutdowns (#47).
- Initial release