v0.7.0: GCE L7 Ingress controller [beta]

Image: gcr.io/google_containers/glbc:0.7.0
Changelog:

• Multi zone support: The Ingress controller now creates an instance group per zone, where the zone is detected from the label on the Kubernetes node.
• Readiness probe health checks: If the endpoint pod exposes a http readiness probe, the Ingress controller adopts it as the health check endpoint instead of forcing users to expose "/".
• Upstream Instance listing bug fix: Upstream cloud provider library had a bug in listing instances in projects with a lot of non-kubernetes vms.

Getting Started:
Kick the tires by creating an ingress. You can also perform a dry run on any machine where kubectl works against a Kubernetes cluster by invoke the release binary with glbc --running-in-cluster=false. A dry run will not create any cloud resources.

Make sure you satisfy the pre-requisites, consult these docs for troubleshooting and see the wishlist if you'd like to contribute to the next release.

v0.6.3: GCE L7 Ingress controller [beta]

Image: gcr.io/google_containers/glbc:0.6.3
Changelog:

• Stability fixes: Improve reliability of GCE client creation/Url map updates etc
• Controller UID: The ingress controller allocates UIDs so 2 kubernetes clusters in the same gce project, with the same ingress, don't collide.
• TLS Certificates: Certificates are updated in GCE when Kubernetes secrets change.
• Firewall rules: The single firewall rule required to allow L7 health checks is created and synced automatically by the controller.

Getting Started:
Kick the tires by creating an ingress. You can also perform a dry run on any machine where kubectl works against a Kubernetes cluster by invoke the release binary with glbc --running-in-cluster=false. A dry run will not create any cloud resources.

Make sure you satisfy the pre-requisites, consult these docs for troubleshooting and see the wishlist if you'd like to contribute to the next release.

v0.6.0: GCE L7 Ingress controller [beta]

Image: gcr.io/google_containers/glbc:0.6.0
Changelog:

• Repository subdirectory renamed to "ingress": A capital 'I' in the directory name is just not right.
• TLS termination: A user specified secret containing TLS certificates is now used for TLS termination.
• Annotation to disallow HTTP: A new annotation called kubernetes.io/ingress.allowHTTP to support pure HTTPS.

Getting Started:
Kick the tires by creating an ingress. You can also perform a dry run on any machine where kubectl works against a Kubernetes cluster by invoke the release binary with glbc --running-in-cluster=false. A dry run will not create any cloud resources.

Make sure you satisfy the pre-requisites, consult these docs for troubleshooting and see the wishlist if you'd like to contribute to the next release.

v0.5.2: GCE L7 Ingress controller [beta]

Image: gcr.io/google_containers/glbc:0.5.2
Changelog:

• All L7s use a single Instance Group: A GCE schema change disallows multiple Instance Groups for the same set of loadbalanced nodes. This change teaches the Ingress controller to add all required Kubernetes NodePorts to a single Instance Group, instead of 1 per port as it had done previously.
• Refactor L7 modules: Split various cloud components into submodules with interfaces, fakes and unittests.
• Fix events: Fixes a bug where the controller reported un-recognized event types to the apiserver.

Getting Started:
Kick the tires by creating an ingress. You can also perform a dry run on any machine where kubectl works against a Kubernetes cluster by invoke the release binary with glbc --running-in-cluster=false. A dry run will not create any cloud resources.

Make sure you satisfy the pre-requisites, consult these docs for troubleshooting and see the wishlist if you'd like to contribute to the next release.

v0.5.1: GCE L7 Ingress controller [beta]

Image: gcr.io/google_containers/glbc:0.5.1

Kick the tires by creating an ingress. You can also perform a dry run on any machine where kubectl works against a Kubernetes cluster by invoke the release binary with glbc --running-in-cluster=false. A dry run will not create any cloud resources.

Make sure you satisfy the pre-requisites, consult these docs for troubleshooting and see the wishlist if you'd like to contribute to the next release.