Container Linux Update Operator Versions Save

• Add reboot windows (#170, thanks @mkoerperriag!)
• Prevent CLUO from marking nodes as schedulable if they were previously marked as unschedulable by another source (#176, thanks @hankjacobs!)
• Restructure examples directory for easier deployment (#175, #178, thanks @laghoule!)

Image is located at quay.io/coreos/container-linux-update-operator:v0.7.0

• Add the container-linux-update.v1.coreos.com/reboot-needed: "true" label to reflect the "reboot-needed" annotation (#168)

Image is located at quay.io/coreos/container-linux-update-operator:v0.6.0

• Add support for Kubernetes v1.9 (#164)

Image is located at quay.io/coreos/container-linux-update-operator:v0.5.0

• Improve agent to wait for pod deletions in parallel (#150)
• Fix update-agent to consider pod UID or name during deletion waits (#151)
  • Handles cases where pods are recreated with the same name
• Fix to return the correct error on D-Bus Call failures (#148)
• Fix missing RBAC get pods permission to examples (#147)

Image is located at quay.io/coreos/container-linux-update-operator:v0.4.1

• Switch leader election to use a ConfigMap instead of an Endpoint (#140)
• Fix update-agent to wait for pods to delete before reboots (#137)
• Add update-agent flag --grace-period to set the maximum time to wait for pods to terminate (#137)
• Deprecate the --analytics flag. Google Analytics is removed (#144)
• Fix log line "Waiting for before-reboot annotations..." (#133)
• Update client-go for Kubernetes 1.7+ (#140)
• Add RBAC ClusterRole, binding, and namespace to examples (#139)

Image is located at quay.io/coreos/container-linux-update-operator:v0.4.0

Fix log message in steady-state that erroneously claimed labels are changing (#126)

Image is located at quay.io/coreos/container-linux-update-operator:v0.3.1

• Add support for before-reboot and after-reboot required annotations (#121)
  • Allow setting update-operator required --before-reboot-annotations and --after-reboot-annotations
  • Add before-reboot label before reboots and add after-reboot label after reboot (but before schedulable)
• Auto-label Container Linux nodes with agent=true for backwards compatibility (#117)
  • Set --auto-label-container-linux=true to enable Container Linux labeling
• Add docs on CLUO labels and annotations (#123)
• Add docs and examples of before-reboot and after-reboot DaemonSet (#124)

Image is located at quay.io/coreos/container-linux-update-operator:v0.3.0

• Deprecate manage-agent behavior. The manage-agent option defaults to false and will be removed in the future. (#98, #76)
• Fix possible deadlock in update-agent if update-engine requests a reboot before we are initialized.
• Add additional logging in update-operator with names of nodes being rebooted (#72)

Image is located at quay.io/coreos/container-linux-update-operator:v0.2.2

• Fix update-operator to deploy update-agent version with a "v" prefix (#71)
• Fix example Kubernetes manifests to use release tags (#73)
• Fix update-agent to retry to avoid exiting on watch termination (#89)
• Reboot nodes, even if a pod fails to be deleted (#89)
• Update to build binaries with Go 1.8.3 (#87)
• Update container image base from alpine:3.4 to alpine:3.6 (#95)

Changes

• Allow operator to manage the associated update agents
• Add version information and flag to components
• Update to client-go 3.0 for Kubernetes 1.6 compatibility