Notable changes since v0.9.6

Full changelog can be seen at v0.9.6...v0.9.7

Please see our roadmap for details on upcoming releases.

Actions required

• #639: Users of the experimental TLS bootstrap feature are required to run the following actions:

1. Run kube-aws render stack to update controller/worker user data templates

2. Move the bootstrap token from ./credentials/tokens.csv to its own file ./credentials/kubelet-tls-bootstrap-token

   # Remove the following line from ./credentials/tokens.csv, and move <token>
   # (with no leading/trailing blank chars and lines) to it's own file
   # ./credentials/kubelet-tls-bootstrap-token
   <token>,kubelet-bootstrap,10001,system:kubelet-bootstrap
   

3. Run kube-aws update to update the cluster. This operation will cause controllers and workers to be replaced

• #629: experimenetal.clusterAutoscalerSupport.enabled was removed in favor of addons.clusterAutoscaler
• #629: worker.nodePools[].clusterAutoscaler.(min|max)Size was removed in favor of worker.nodePools[].autoscaling.clusterAutoscaler.enabled

Known issues

• etcdDataVolumeEphemeral is broken. Please don't turn it until it is fixed and keep using EBS based data volumes instead. #446

Features

• Kubernetes 1.6.3
  • #624: Update default dashboard to 1.6.0 (Thanks to @Vrtak-CZ)
  • #646: Update versions for various components.(Thanks to @camilb)
  • #659: Update Kubernetes dashboard to v1.6.1. Update calico to v2.2.1.(Thanks to @camilb)
• #559: Add sprig templating functions (Thanks to @tyrannasaurusbanks)
• #607: More flexible configuration for IAM and stable naming for roles (Thanks to @Fsero)
• #629: Re: cluster-autoscaler support(Thanks to @redbaron for reviewing)
  • Follow-ups
    • #724: Add critical-pod annotation to CA
    • #738: Add CA a CriticalAddonsOnly toleration
• #645: NVIDIA driver installation support on GPU instances(Thanks to @everpeace)
• #707: Send Journald logs to AWS CloudWatch(Thanks to @jollinshead)
• #737: Drain node on spot instance termination notice as well(Thanks to @danielfm)

Enhancements

• #615: kube-dns improvements(Thanks to @danielfm)
• #616: Improve taints validation(Thanks to @danielfm)
• #618: RBAC setup improvements(Thanks to @danielfm)
• #625: Fix dashboard version in labels and remove version from RC name (Thanks to @Vrtak-CZ)
• #650: Label masters with 'node-role.kubernetes.io/master' label(Thanks to @redbaron)
• #652: Label masters with 'node-role.kubernetes.io/master' label(Thanks to @redbaron)
• #663: Make kubelet flags more consistent(Thanks to @redbaron)
• #693: Spot fleet detailed monitoring(Thanks to @paalkr)
• #682: Allow userdata to be split across multiple parts(Thanks to @redbaron)
• #701: core: update cluster-proportional-autoscaler to v1.1.2(Thanks to @harsha-y)
• #718: Report errors if empty nodepool is specified(Thanks to @redbaron)

Fixes

• #619: Fix IamFleetRole syntax (Thanks to @danielfm)
• #632: Fix unable to backup namespaced resources(Thanks to @cheungpat)
• #634: Fixes leading slash on s3:prefix removed to fix ListObject permission denied error during etcdadm save(Thanks to @jeremyd)
• #639: Avoid unnecessary node replacements when TLS bootstrapping is enabled(Thanks to @danielfm)
• #641: adding --rm to docker run aws_cli commands in etcdadm(Thanks to @jeremyd)
• #657: Fix typo in help message(Thanks to @ytsarev)
• #662: Fix taint being assigned as labels(Thanks to @redbaron)
• #671: Fix "install-kube-system" script when "clusterAutoscaler" is disabled(Thanks to @camilb)
• #674: Improved node drainer(Thanks to @danielfm)
• #676: Ensure TLSBootstrapToken related code is not written if TLSBootstrap is disabled(Thanks to @jollinshead)
• #687: calico-node - DaemonSet tolerating all taints(Thanks to @jeffersongirao)
• #686: Heapster service account and cluster role binding(Thanks to @c-knowles)
  Follow-up: #692: Apply heapster RBAC setup properly(Thanks to @c-knowles)
• #695: core: Add ListBucket explicitly to EtcdSnapshotsS3Bucket(Thanks to @trinitronx)
• #697: Fix the issue that cluster-autoscaler never scale-down the cluster
• #705: Use docker instead of rkt for regular etcdadm tasks(Thanks to @ytsarev)
• #710: Fix node drain error when trying to evict pods from jobs(Thanks to @danielfm)
• #711: Remove unused sysctl override for nf_conntrack_max(Thanks to @danielfm)
• #722: Additional propagation of etcd version for etcdadm(Thanks to @ytsarev)
• #735: Inject stack name into userdata for nodepool workers(Thanks to @redbaron)

Documentation

• #626: Fix quote on "Launch the secondary node"(Thanks to @velo)
• #635: cluster.yaml: Additional note for etcd.count
• #637: Fix destroy doc(Thanks to @jorge07)
• #689: added required --s3-url parameter(Thanks to @spatronis)
• #696: Introduce kube-aws slack channel in README
• #700: Enhance cluster.yaml documentation(Thanks to @ytsarev)

Refactoring

• #680: Remove ancient CoreOS version check which is not relevant anymore(Thanks to @redbaron)
• #678: Kill dead code and restore DNS config validation(Thanks to @redbaron)
• #670: Remove obsolete etcd locking logic(Thanks to @redbaron)

Testing

• #669: Make go test timeout longer enough for Travis
• #723: e2e: Fix flag provided but not defined: -check_version_skew error

Project

• #739: Update OWNERS