Changelog
Changelog for EKS Anywhere releases
Announcements
- EKS Anywhere release
v0.19.0 introduces support for creating Kubernetes version v1.29 clusters. A conformance test was promoted in Kubernetes v1.29 that verifies that Services serving different L4 protocols with the same port number can co-exist in a Kubernetes cluster. This is not supported in Cilium, the CNI deployed on EKS Anywhere clusters, because Cilium currently does not differentiate between TCP and UDP protocols for Kubernetes Services. Hence EKS Anywhere v1.29 and above clusters will not pass this specific conformance test. This service protocol differentiation is being tracked in an upstream Cilium issue and will be supported in a future Cilium release. A future release of EKS Anywhere will include the patched Cilium version when it is available.
Refer to the following links for more information regarding the conformance test:
- The Bottlerocket project will not be releasing bare metal variants for Kubernetes versions v1.29 and beyond. Hence Bottlerocket is not a supported operating system for creating EKS Anywhere bare metal clusters with Kubernetes versions v1.29 and above. However, Bottlerocket is still supported for bare metal clusters running Kubernetes versions v1.28 and below.
Refer to the following links for more information regarding the deprecation:
- On January 31, 2024, a High-severity vulnerability CVE-2024-21626 was published affecting all
runc versions <= v1.1.11. This CVE has been fixed in runc version v1.1.12, which has been included in EKS Anywhere release v0.18.6. In order to fix this CVE in your new/existing EKS-A cluster, you MUST build or download new OS images pertaining to version v0.18.6 and create/upgrade your cluster with these images.
Refer to the following links for more information on the steps to mitigate the CVE.
- On October 11, 2024, a security issue CVE-2024-9594 was discovered in the Kubernetes Image Builder where default credentials are enabled during the image build process when using the Nutanix, OVA, QEMU or raw providers. The credentials can be used to gain root access. The credentials are disabled at the conclusion of the image build process. Kubernetes clusters are only affected if their nodes use VM images created via the Image Builder project. Clusters using virtual machine images built with Kubernetes Image Builder
version v0.1.37 or earlier are affected if built with the Nutanix, OVA, QEMU or raw providers. These images built using previous versions of image-builder will be vulnerable only during the image build process, if an attacker was able to reach the VM where the image build was happening, login using these default credentials and modify the image at the time the image build was occurring. This CVE has been fixed in image-builder versions >= v0.1.38, which has been included in EKS Anywhere releases v0.19.11 and v0.20.8.
General Information
- When upgrading to a new minor version, a new OS image must be created using the new image-builder CLI pertaining to that release.
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- cilium:
v1.13.20-eksa.1 to v1.13.21-eksa.5
- cloud-provider-vsphere
v1.29.1 to v1.29.2v1.30.1 to v1.30.2
- EKS Distro:
- cluster-api-provider-vsphere(CAPV):
v1.10.3 to v1.10.4
- etcdadm-bootstrap-provider:
v1.0.14 to v1.0.15
- kube-vip:
v0.8.4 to v0.8.6
Fixed
- Release init-lock when the owner machine fails to launch. (#41
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- EKS Distro:
- image-builder:
v0.1.36 to v0.1.39
- cluster-api-provider-vsphere(CAPV):
v1.10.3 to v1.10.4
- etcdadm-controller:
v1.0.23 to v1.0.24
- etcdadm-bootstrap-provider:
v1.0.13 to v1.0.14
- kube-vip:
v0.8.3 to v0.8.4
- containerd:
v1.7.22 to v1.7.23
- runc:
v1.1.14 to v1.1.15
- local-path-provisioner:
v0.0.29 to v0.0.30
Fixed
- Skip hardware validation logic for InPlace upgrades. #8779
- Status reconciliation of etcdadm cluster in etcdadm-controller when etcd-machines are unhealthy. #63
- Skip generating AWS IAM Kubeconfig on cluster upgrade. #8851
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- EKS Distro:
- cilium:
v1.13.19 to v1.13.20
- image-builder:
v0.1.30 to v0.1.36
- cluster-api-provider-vsphere(CAPV):
v1.10.2 to v1.10.3
Fixed
- Fixed support for efi on rhel 9 raw builds. (#3824
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- EKS Distro:
- cilium:
v1.13.18 to v1.13.19
- containerd
v1.7.21 to v.1.7.22
- etcdadm-controller:
v1.0.22 to v1.0.23
- kube-vip:
v0.8.2 to v0.8.3
- Kube-rbac-proxy:
v0.18.0 to v0.18.1
Added
- Enable EFI boot support on RHEL9 images for bare-meal. (#3684
)
Fixed
- Status reconciliation of etcdadm cluster in etcdadm-controller when
etcd-machines are unhealthy. (#63
)
- Skip hardware validation logic for InPlace upgrades. ([#8779]https://github.com/aws/eks-anywhere/pull/8779))
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- EKS Distro:
- Tinkerbell Stack:
- runc
v1.1.13 to v1.1.14
- containerd
v1.7.20 to v.1.7.21
- local-path-provisioner
v0.0.28 to v0.0.29
Fixed
- Rollout new nodes for OSImageURL change on spec without changing K8s version (#8656
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
- EKS Distro:
- EKS Anywhere Packages Controller:
v0.4.3 to v0.4.4
- Helm:
v3.15.3 to v3.15.4
Fixed
- Fix Kubelet Configuration apply when host OS config is specified. (#8606
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
— |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
Changed
Added
- Enable Tinkerbell stack to use dhcprelay instead of using smee in hostNetwork mode. (#8568
)
Fixed
- Enable
lb_class_only env var on kube-vip so that it only manages IP for services with LoadBalancerClass set to kube-vip.io/kube-vip-class on the service. (#8493
)
- Nil pointer panic for etcdadm-controller when
apiserver-etcd-client secret got deleted. (#62
)
Changed
- kube-vip:
v0.8.1 to v0.8.2
- cilium:
v1.13.16 to v1.13.18
- cert-manager:
v1.14.5 to v1.14.7
- etcdadm-controller:
v1.0.21 to v1.0.22
- local-path-provisioner:
v0.0.27 to v0.0.28
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.0 |
✔ |
— |
— |
— |
— |
| Bottlerocket 1.19.4 |
— |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
Fixed
- Fix panic when datacenter obj is not found (8495
)
- Fix Subnet Validation Bug for Nutanix provider (8499
)
- Fix machine config panic when ref object not found (8533
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.0 |
✔ |
— |
— |
— |
— |
| Bottlerocket 1.19.4 |
— |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
Fixed
- Fix cluster status reconciliation for control plane and worker nodes (8455
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.0 |
✔ |
— |
— |
— |
— |
| Bottlerocket 1.19.4 |
— |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
✔ |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Added
- Support for Kubernetes v1.30
- Support for configuring kube-apiserver flags in cluster spec (#7755
)
- Redhat 9 support for Bare Metal (#3032
)
- Support for configuring kubelet settings in cluster spec (#7708
)
- Support for control plane failure domains on Nutanix (#8192
)
Changed
- Generate cluster config command includes OSImageURL in tinkerbell machine config objects (#8226
)
- Added EKS-D for 1-30:
- Cilium:
v1.13.9 to v1.13.16
- Cluster API:
v1.6.1 to v1.7.2
- Cluster API Provider vSphere:
v1.8.5 to v1.10.0
- Cluster API Provider Nutanix:
v1.2.3 to v1.3.5
- Flux:
v2.2.3 to v2.3.0
- Kube-vip:
v0.7.0 to v0.8.0
- Image-builder:
v0.1.24 to v0.1.26
- Kind:
v0.22.0 to v0.23.0
- Etcdadm Controller:
v1.0.17 to v1.0.21
- Tinkerbell Stack:
- Cluster API Provider Tinkerbell:
v0.4.0 to v0.5.3
- Hegel:
v0.10.1 to v0.12.0
- Rufio:
afd7cd82fa08dae8f9f3ffac96eb030176f3abbd to v0.3.3
- Tink:
v0.8.0 to v0.10.0
- Boots/Smee:
v0.8.1 to v0.11.0
- Hook:
9d54933a03f2f4c06322969b06caa18702d17f66 to v0.8.1
- Charts:
v0.4.5
Note: The Boots service has been renamed to Smee by the upstream tinkerbell community with this version upgrade. Any reference to Boots or Smee in our docs refer to the same service.
Removed
- Support for Kubernetes v1.25
- Support for certain curated packages CLI commands (#8240
)
Fixed
- CLI commands for packages to honor the registry mirror setup in cluster spec (#8026
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.5 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Upgraded
- EKS Distro:
- Image-builder:
v0.1.36 to v0.1.39 (CVE-2024-9594
)
- containerd:
v1.7.22 to v1.7.23
- Cilium:
v1.13.19 to v1.13.20
- etcdadm-controller:
v1.0.23 to v1.0.24
- etcdadm-bootstrap-provider:
v1.0.13 to v1.0.14
- local-path-provisioner:
v0.0.29 to v0.0.30
- runc:
v1.1.14 to v1.1.15
Fixed
- Skip hardware validation logic for InPlace upgrades. #8779
- Status reconciliation of etcdadm cluster in etcdadm-controller when etcd-machines are unhealthy. #63
- Skip generating AWS IAM Kubeconfig on cluster upgrade. #8851
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.20.0 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Upgraded
- EKS Distro:
- EKS Anywhere Packages:
v0.4.3 to v0.4.4
- Cilium:
v1.13.18 to v1.13.19
- containerd:
v1.7.20 to v1.7.22
- runc:
v1.1.13 to v1.1.14
- local-path-provisioner:
v0.0.28 to v0.0.29
- etcdadm-controller:
v1.0.22 to v1.0.23
- New base images with CVE fixes for Amazon Linux 2
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Upgraded
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Upgraded
Changed
- Added additional validation before marking controlPlane and workers ready #8455
Fixed
- Fix panic when datacenter obj is not found #8494
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Upgraded
- Cluster API Provider Nutanix:
v1.3.3 to v1.3.5
- Image Builder:
v0.1.24 to v0.1.26
- EKS Distro:
Changed
- Updated cluster status reconciliation logic for worker node groups with autoscaling
configuration #8254
- Added logic to apply new hardware on baremetal cluster upgrades #8288
Fixed
- Fixed bug when installer does not create CCM secret for Nutanix workload cluster #8191
- Fixed upgrade workflow for registry mirror certificates in EKS Anywhere packages #7114
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
- Backporting dependency bumps to fix vulnerabilities #8118
- Upgraded EKS-D:
Fixed
- Fixed cluster directory being created with root ownership #8120
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
- Upgraded EKS-Anywhere Packages from
v0.4.2 to v0.4.3
Fixed
- Fixed registry mirror with authentication for EKS Anywhere packages
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
- Support Docs site for penultime EKS-A version #8010
- Update Ubuntu 22.04 ISO URLs to latest stable release #3114
- Upgraded EKS-D:
Fixed
- Added processor for Tinkerbell Template Config #7816
- Added nil check for eksa-version when setting etcd url #8018
- Fixed registry mirror secret credentials set to empty #7933
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
- Updated helm to v3.14.3 #3050
Fixed
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
- Update CAPC to 0.4.10-rc1 #3105
- Upgraded EKS-D:
Fixed
- Fixing tinkerbell action image URIs while using registry mirror with proxy cache.
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.2 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Changed
Added
- Preflight check for upgrade management components such that it ensures management components is at most 1 EKS Anywhere minor version greater than the EKS Anywhere version of cluster components #7800
.
Fixed
- EKS Anywhere package bundles
ending with 152, 153, 154, 157 have image tag issues which have been resolved in bundle 158. Example for kubernetes version v1.29 we have
public.ecr.aws/eks-anywhere/eks-anywhere-packages-bundles:v1-29-158
- Fixed InPlace custom resources from being created again after a successful node upgrade due to delay in objects in client cache #7779
.
- Fixed #7623
by encoding the basic auth credentials to base64 when using them in templates #7829
.
- Added a fix
for error that may occur during upgrading management components where if the cluster object is modified by another process before applying, it throws the conflict error prompting a retry.
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.0 |
✔ |
* |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
✔ |
— |
* EKS Anywhere issue regarding deprecation of Bottlerocket bare metal variants
Added
- Support for Kubernetes v1.29
- Support for in-place EKS Anywhere and Kubernetes version upgrades on Bare Metal clusters
- Support for horizontally scaling
etcd count in clusters with external etcd deployments (#7127
)
- External
etcd support for Nutanix (#7550
)
- Etcd encryption for Nutanix (#7565
)
- Nutanix Cloud Controller Manager integration (#7534
)
- Enable image signing for all images used in cluster operations
- RedHat 9 support for CloudStack (#2842
)
- New
upgrade management-components command which upgrades management components independently of cluster components (#7238
)
- New
upgrade plan management-components command which provides new release versions for the next management components upgrade (#7447
)
- Make
maxUnhealthy count configurable for control plane and worker machines (#7281
)
Changed
- Unification of controller and CLI workflows for cluster lifecycle operations such as create, upgrade, and delete
- Perform CAPI Backup on workload cluster during upgrade(#7364
)
- Extend
maxSurge and maxUnavailable configuration support to all providers
- Upgraded Cilium to v1.13.19
- Upgraded EKS-D:
- Cluster API Provider AWS Snow:
v0.1.26 to v0.1.27
- Cluster API:
v1.5.2 to v1.6.1
- Cluster API Provider vSphere:
v1.7.4 to v1.8.5
- Cluster API Provider Nutanix:
v1.2.3 to v1.3.1
- Flux:
v2.0.0 to v2.2.3
- Kube-vip:
v0.6.0 to v0.7.0
- Image-builder:
v0.1.19 to v0.1.24
- Kind:
v0.20.0 to v0.22.0
Removed
Fixed
- Validate OCI namespaces for registry mirror on Bottlerocket (#7257
)
- Make Cilium reconciler use provider namespace when generating network policy (#7705
)
- EKS Anywhere v0.18.7 Admin AMI with CVE fixes for Amazon Linux 2
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.0 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
- EKS Anywhere v0.18.6 Admin AMI with CVE fixes for
runc
- New base images with CVE fixes for Amazon Linux 2
- Bottlerocket
v1.15.1 to 1.19.0
- runc
v1.1.10 to v1.1.12 (CVE-2024-21626
)
- containerd
v1.7.11 to v.1.7.12
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.19.0 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.x |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
- New EKS Anywhere Admin AMI with CVE fixes for Amazon Linux 2
- New base images with CVE fixes for Amazon Linux 2
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.15.1 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
Feature
- Nutanix: Enable api-server audit logging for Nutanix (#2664
)
Bug
- CNI reconciler now properly pulls images from registry mirror instead of public ECR in airgapped environments: #7170
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.15.1 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
Fixed
- Etcdadm: Renew client certificates when nodes rollover (etcdadm/#56
)
- Include DefaultCNIConfigured condition in Cluster Ready status except when Skip Upgrades is enabled (#7132
)
- EKS Distro (Kubernetes):
v1.25.15 to v1.25.16v1.26.10 to v1.26.11v1.27.7 to v1.27.8v1.28.3 to v1.28.4
- Etcdadm Controller:
v1.0.15 to v1.0.16
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.15.1 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
Fixed
- Image Builder: Correctly parse
no_proxy inputs when both Red Hat Satellite and Proxy is used in image-builder. (#2664
)
- vSphere: Fix template tag validation by specifying the full template path (#6437
)
- Bare Metal: Skip
kube-vip deployment when TinkerbellDatacenterConfig.skipLoadBalancerDeployment is set to true. (#6990
)
Other
- Security: Patch incorrect conversion between uint64 and int64 (#7048
)
- Security: Fix incorrect regex for matching curated package registry URL (#7049
)
- Security: Patch malicious tarballs directory traversal vulnerability (#7057
)
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.15.1 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
- EKS Distro (Kubernetes):
v1.25.14 to v1.25.15v1.26.9 to v1.26.10v1.27.6 to v1.27.7v1.28.2 to v1.28.3
- Etcdadm Bootstrap Provider:
v1.0.9 to v1.0.10
- Etcdadm Controller:
v1.0.14 to v1.0.15
- Cluster API Provider CloudStack:
v0.4.9-rc7 to v0.4.9-rc8
- EKS Anywhere Packages Controller :
v0.3.12 to v0.3.13
Bug
- Bare Metal: Ensure the Tinkerbell stack continues to run on management clusters when worker nodes are scaled to 0 (#2624
)
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.15.1 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
✔ |
✔ |
— |
| RHEL 9.x |
— |
— |
✔ |
— |
— |
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.15.1 |
1.15.1 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
9.x, 8.7 |
8.7 |
Not supported |
Added
- Etcd encryption for CloudStack and vSphere: #6557
- Generate TinkerbellTemplateConfig command: #3588
- Support for modular Kubernetes version upgrades with bare metal: #6735
- OSImageURL added to Tinkerbell Machine Config
- Bare metal out-of-band webhook: #5738
- Support for Kubernetes v1.28
- Support for air gapped image building: #6457
- Support for RHEL 8 and RHEL 9 for Nutanix provider: #6822
- Support proxy configuration on Redhat image building #2466
- Support Redhat Satellite in image building #2467
Changed
- KinD-less upgrades: #6622
- Management cluster upgrades don’t require a local bootstrap cluster anymore.
- The control plane of management clusters can be upgraded through the API. Previously only changes to the worker nodes were allowed.
- Increased control over upgrades by separating external etcd reconciliation from control plane nodes: #6496
- Upgraded Cilium to 1.12.15
- Upgraded EKS-D:
- Cluster API Provider CloudStack:
v0.4.9-rc6 to v0.4.9-rc7
- Cluster API Provider AWS Snow:
v0.1.26 to v0.1.27
- Upgraded CAPI to
v1.5.2
Removed
- Support for Kubernetes v1.23
Fixed
- Fail on
eksctl anywhere upgrade cluster plan -f: #6716
- Error out when management kubeconfig is not present for workload cluster operations: 6501
- Empty vSphereMachineConfig users fails CLI upgrade: 5420
- CLI stalls on upgrade with Flux Gitops: 6453
Bug
- CNI reconciler now properly pulls images from registry mirror instead of public ECR in airgapped environments: #7170
- Waiting for control plane to be fully upgraded: #6764
Other
- Check for k8s version in the Cloudstack template name: #7130
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.14.3 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
_ |
✔ |
— |
- Cluster API Provider CloudStack:
v0.4.9-rc7 to v0.4.9-rc8
Supported Operating Systems
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu 20.04 |
✔ |
✔ |
✔ |
— |
✔ |
| Ubuntu 22.04 |
✔ |
✔ |
✔ |
— |
— |
| Bottlerocket 1.14.3 |
✔ |
✔ |
— |
— |
— |
| RHEL 8.7 |
✔ |
✔ |
_ |
✔ |
— |
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.14.3 |
1.14.3 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Added
- Enabled audit logging for
kube-apiserver on baremetal provider (#6779
).
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.14.3 |
1.14.3 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Fixed
- Fixed cli upgrade mgmt kubeconfig flag (#6666
)
- Ignore node taints when scheduling Cilium preflight daemonset (#6697
)
- Baremetal: Prevent bare metal machine config references from changing to existing machine configs (#6674
)
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.14.0 |
1.14.0 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Fixed
- Bare Metal: Ensure new worker node groups can reference new machine configs (#6615
)
- Bare Metal: Fix
writefile action to ensure Bottlerocket configs write content or error (#2441
)
Added
- Added support for configuring healthchecks on EtcdadmClusters using
etcdcluster.cluster.x-k8s.io/healthcheck-retries annotation (aws/etcdadm-controller#44
)
- Add check for making sure quorum is maintained before deleting etcd machines (aws/etcdadm-controller#46
)
Changed
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.14.0 |
1.14.0 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Fixed
- Fix worker node groups being rolled when labels adjusted #6330
- Fix worker node groups being rolled out when taints are changed #6482
- Fix vSphere template tags validation to run on the control plane and etcd
VSpherMachinesConfig #6591
- Fix Bare Metal upgrade with custom pod CIDR #6442
Added
- Add validation for missing management cluster kubeconfig during workload cluster operations #6501
Supported OS version details
|
vSphere |
Bare Metal |
Nutanix |
CloudStack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
|
22.04 |
22.04 |
22.04 |
Not supported |
Not supported |
| Bottlerocket |
1.14.0 |
1.14.0 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Note: We have updated the image-builder docs to include the latest enhancements. Please refer to the image-builder docs
for more details.
Added
- Add support for AWS CodeCommit repositories in FluxConfig with git configuration #4290
- Add new information to the EKS Anywhere Cluster status #5628
:
- Add the
ControlPlaneInitialized, ControlPlaneReady, DefaultCNIConfigured, WorkersReady, and Ready conditions.
- Add the
observedGeneration field.
- Add the
failureReason field.
- Add support for different machine templates for control plane, etcd, and worker node in vSphere provider #4255
- Add support for different machine templates for control plane, etcd, and worker node in Cloudstack provider #6291
- Add support for Kubernetes version 1.25, 1.26, and 1.27 to CloudStack provider #6167
- Add bootstrap cluster backup in the event of cluster upgrade error #6086
- Add support for organizing virtual machines into categories with the Nutanix provider #6014
- Add support for configuring
egressMasqueradeInterfaces option in Cilium CNI via EKS Anywhere cluster spec #6018
- Add support for a flag for create and upgrade cluster to skip the validation
--skip-validations=vsphere-user-privilege
- Add support for upgrading control plane nodes separately from worker nodes for vSphere, Nutanix, Snow, and Cloudstack providers #6180
- Add preflight validation to prevent skip eks-a minor version upgrades #5688
- Add preflight check to block using kindnetd CNI in all providers except Docker [#6097]https://github.com/aws/eks-anywhere/issues/6097
- Added feature to configure machine health checks for API managed clusters and a new way to configure health check timeouts via the EKKSA spec. [#6176]https://github.com/aws/eks-anywhere/pull/6176
Upgraded
- Cluster API Provider vSphere:
v1.6.1 to v1.7.0
- Cluster API Provider Cloudstack:
v0.4.9-rc5 to v0.4.9-rc6
- Cluster API Provider Nutanix:
v1.2.1 to v1.2.3
Cilium Upgrades
-
Cilium: v1.11.15 to v1.12.11
Note: If you are using the vSphere provider with the Redhat OS family, there is a known issue with VMWare and the new Cilium version that only affects our Redhat variants. To prevent this from affecting your upgrade from EKS Anywhere v0.16 to v0.17, we are adding a temporary daemonset to disable UDP offloading on the nodes before upgrading Cilium. After your cluster is upgraded, the daemonset will be deleted. This note is strictly informational as this change requires no additional effort from the user.
Changed
- Change the default node startup timeout from 10m to 20m in Bare Metal provider #5942
- EKS Anywhere now fails on pre-flights if a user does not have required permissions. #5865
eksaVersion field in the cluster spec is added for better representing CLI version and dependencies in EKS-A cluster #5847
- vSphere datacenter insecure and thumbprint is now mutable for upgrades when using full lifecycle API [6143]https://github.com/aws/eks-anywhere/issues/6143
Fixed
- Fix cluster creation failure when the
<Provider>DatacenterConfig is missing apiVersion field #6096
- Allow registry mirror configurations to be mutable for Bottlerocket OS #2336
- Patch an issue where mutable fields in the EKS Anywhere CloudStack API failed to trigger upgrades #5910
- image builder: Fix runtime issue with git in image-builder v0.16.2 binary #2360
- Bare Metal: Fix issue where metadata requests that return non-200 responses were incorrectly treated as OK #2256
Known Issues:
- Upgrading Docker clusters from previous versions of EKS Anywhere may not work on Linux hosts due to an issue in the Cilium 1.11 to 1.12 upgrade. Docker clusters is meant solely for testing and not recommended or support for production use cases. There is currently no fixed planned.
- If you are installing EKS Anywhere Packages, Kubernetes versions 1.23-1.25 are incompatible with Kubernetes versions 1.26-1.27 due to an API difference. This means that you may not have worker nodes on Kubernetes version <= 1.25 when the control plane nodes are on Kubernetes version >= 1.26. Therefore, if you are upgrading your control plane nodes to 1.26, you must upgrade all nodes to 1.26 to avoid failures.
- There is a known bug
with systemd >= 249 and all versions of Cilium. This is currently known to only affect Ubuntu 22.04. This will be fixed in future versions of EKS Anywhere. To work around this issue, run one of the follow options on all nodes.
Option A
# Does not persist across reboots.
sudo ip rule add from all fwmark 0x200/0xf00 lookup 2004 pref 9
sudo ip rule add from all fwmark 0xa00/0xf00 lookup 2005 pref 10
sudo ip rule add from all lookup local pref 100
Option B
# Does persist across reboots.
# Add these values /etc/systemd/networkd.conf
[Network]
ManageForeignRoutes=no
ManageForeignRoutingPolicyRules=no
Deprecated
- The bundlesRef field in the cluster spec is now deprecated in favor of the new
eksaVersion field. This field will be deprecated in three versions.
Removed
- Installing vSphere CSI Driver as part of vSphere cluster creation. For more information on how to self-install the driver refer to the documentation here
⚠️ Breaking changes
- CLI:
--force-cleanup has been removed from create cluster, upgrade cluster and delete cluster commands. For more information on how to troubleshoot issues with the bootstrap cluster refer to the troubleshooting guide (1
and 2
). #6384
Changed
- Bump up the worker count for etcdadm-controller from 1 to 10 #34
- Add 2X replicas hard limit for rolling out new etcd machines #37
Fixed
- Fix code panic in healthcheck loop in etcdadm-controller #41
- Fix deleting out of date machines in etcdadm-controller #40
Fixed
- Fix support for having management cluster and workload cluster in different namespaces #6414
Changed
- During management cluster upgrade, if the backup of CAPI objects of all workload clusters attached to the management cluster fails before upgrade starts, EKS Anywhere will only backup the management cluster #6360
- Update kubectl wait retry policy to retry on TLS handshake errors #6373
Removed
- Removed the validation for checking management cluster bundle compatibility on create/upgrade workload cluster #6365
Fixes
- CLI: Ensure importing packages and bundles honors the insecure flag #6056
- vSphere: Fix credential configuration when using the full lifecycle controller #6058
- Bare Metal: Fix handling of Hardware validation errors in Tinkerbell full lifecycle cluster provisioning #6091
- Bare Metal: Fix parsing of bare metal cluster configurations containing embedded PEM certs #6095
Upgrades
- AWS Cloud Provider: v1.27.0 to v1.27.1
- EKS Distro:
- Kubernetes v1.24.13 to v1.24.15
- Kubernetes v1.25.9 to v1.25.11
- Kubernetes v1.26.4 to v1.26.6
- Kubernetes v1.27.1 to v1.27.3
- Cluster API Provider Snow: v0.1.25 to v0.1.26
Added
- Workload clusters full lifecycle API support for CloudStack provider (#2754
)
- Enable proxy configuration for Bare Metal provider (#5925
)
- Kubernetes 1.27 support (#5929
)
- Support for upgrades for clusters with pod disruption budgets (#5697
)
- BottleRocket network config uses mac addresses instead of interface names for configuring interfaces for the Bare Metal provider (#3411
)
- Allow users to configure additional BottleRocket settings
- kernel sysctl settings (#5304
)
- boot kernel parameters (#5359
)
- custom trusted cert bundles (#5625
)
- Add support for IRSA on Nutanix (#5698
)
- Add support for aws-iam-authenticator on Nutanix (#5698
)
- Enable proxy configuration for Nutanix (#5779
)
Upgraded
- Management cluster upgrades will only move management cluster’s components to bootstrap cluster and back. (#5914
)
Fixed
- CloudStack control plane host port is only defaulted in CAPI objects if not provided. (#5792
) (#5736
)
Deprecated
- Add warning to deprecate disableCSI through CLI (#5918
). Refer to the deprecation section
in the vSphere provider documentation for more information.
Removed
Fixed
- Add validation for tinkerbell ip for workload cluster to match management cluster (#5798
)
- Update datastore usage validation to account for space that will free up during upgrade (#5524
)
- Expand GITHUB_TOKEN regex to support fine-grained access tokens (#5764
)
- Display the timeout flags in CLI help (#5637
)
Added
- Added bundles-override to package cli commands (#5695
)
Fixed
Supported OS version details
|
vSphere |
Baremetal |
Nutanix |
Cloudstack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
| Bottlerocket |
1.13.1 |
1.13.1 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Added
- Support for no-timeouts to more EKS Anywhere operations (#5565
)
Changed
- Use kubectl for kube-proxy upgrader calls (#5609
)
Fixed
- Fixed the failure to delete a Tinkerbell workload cluster due to an incorrect SSH key update during reconciliation (#5554
)
- Fixed
machineGroupRef updates for CloudStack and Vsphere (#5313
)
Supported OS version details
|
vSphere |
Baremetal |
Nutanix |
Cloudstack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
| Bottlerocket |
1.13.1 |
1.13.1 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Added
Upgraded
- Cilium updated from version
v1.11.10 to version v1.11.15
Fixed
- Fix http client in file reader to honor the provided HTTP_PROXY, HTTPS_PROXY and NO_PROXY env variables (#5488
)
Supported OS version details
|
vSphere |
Baremetal |
Nutanix |
Cloudstack |
Snow |
| Ubuntu |
20.04 |
20.04 |
20.04 |
Not supported |
20.04 |
| Bottlerocket |
1.13.1 |
1.13.1 |
Not supported |
Not supported |
Not supported |
| RHEL |
8.7 |
8.7 |
Not supported |
8.7 |
Not supported |
Added
- Workload clusters full lifecycle API support for Bare Metal provider (#5237
)
- IRSA support for Bare Metal (#4361
)
- Support for mixed disks within the same node grouping for BareMetal clusters (#3234
)
- Workload clusters full lifecycle API support for Nutanix provider (#5190
)
- OIDC support for Nutanix (#4711
)
- Registry mirror support for Nutanix (#5236
)
- Support for linking EKS Anywhere node VMs to Nutanix projects (#5266
)
- Add
CredentialsRef to NutanixDatacenterConfig to specify Nutanix credentials for workload clusters (#5114
)
- Support for taints and labels for Nutanix provider (#5172
)
- Support for InsecureSkipVerify for RegistryMirrorConfiguration across all providers. Currently only supported for Ubuntu and RHEL OS. (#1647
)
- Support for configuring of Bottlerocket settings. (#707
)
- Support for using a custom CNI (#5217
)
- Ability to configure NTP servers on EKS Anywhere nodes for vSphere and Tinkerbell providers (#4760
)
- Support for nonRootVolumes option in SnowMachineConfig (#5199
)
- Validate template disk size with vSphere provider using Bottlerocket (#1571
)
- Allow users to specify
cloneMode for different VSphereMachineConfig (#4634
)
- Validate management cluster bundles version is the same or newer than bundle version used to upgrade a workload cluster(#5105
)
- Set hostname for Bottlerocket nodes (#3629
)
- Curated Package controller as a package (#831
)
- Curated Package Credentials Package (#829
)
- Enable Full Cluster Lifecycle for curated packages (#807
)
- Curated Package Controller Configuration in Cluster Spec (#5031
)
Upgraded
- Bottlerocket upgraded from
v1.13.0 to v1.13.1
- Upgrade EKS Anywhere admin AMI to Kernel 5.15
- Tinkerbell stack upgraded (#3233
):
- Cluster API Provider Tinkerbell
v0.4.0
- Hegel
v0.10.1
- Rufio
v0.2.1
- Tink
v0.8.0
- Curated Package Harbor upgraded from
2.5.1 to 2.7.1
- Curated Package Prometheus upgraded from
2.39.1 to 2.41.0
- Curated Package Metallb upgraded from
0.13.5 to 0.13.7
- Curated Package Emissary upgraded from
3.3.0 to 3.5.1
Fixed
- Applied a patch that fixes vCenter sessions leak (#1767
)
Breaking changes
- Removed support for Kubernetes 1.21
Fixed
- Fix clustermanager no-timeouts option (#5445
)
Fixed
- Fix kubectl get call to point to full API name (#5342
)
- Expand all kubectl calls to fully qualified names (#5347
)
Added
--no-timeouts flag in create and upgrade commands to disable timeout for all wait operations- Management resources backup procedure with clusterctl
Added
--aws-region flag to copy packages command.
Upgraded
- CAPAS from
v0.1.22 to v0.1.24.
Added
- Enabled support for Kubernetes version 1.25
Added
- support for authenticated pulls from registry mirror (#4796
)
- option to override default nodeStartupTimeout in machine health check (#4800
)
- Validate control plane endpoint with pods and services CIDR blocks(#4816
)
Fixed
- Fixed a issue where registry mirror settings weren’t being applied properly on Bottlerocket nodes for Tinkerbell provider
Added
- Add support for EKS Anywhere on AWS Snow (#1042
)
- Static IP support for BottleRocket (#4359
)
- Add registry mirror support for curated packages
- Add copy packages command (#4420
)
Fixed
- Improve management cluster name validation for workload clusters
Added
- Multi-region support for all supported curated packages
Fixed
- Fixed nil pointer in
eksctl anywhere upgrade plan command
Added
- Workload clusters full lifecycle API support for vSphere and Docker (#1090
)
- Single node cluster support for Bare Metal provider
- Cilium updated to version
v1.11.10
- CLI high verbosity log output is automatically included in the support bundle after a CLI
cluster command error (#1703
implemented by #4289
)
- Allow to configure machine health checks timeout through a new flag
--unhealthy-machine-timeout (#3918
implemented by #4123
)
- Ability to configure rolling upgrade for Bare Metal and Cloudstack via
maxSurge and maxUnavailable parameters
- New Nutanix Provider
- Workload clusters support for Bare Metal
- VM Tagging support for vSphere VM’s created in the cluster (#4228
)
- Support for new curated packages:
- Updated curated packages' versions:
- ADOT
v0.23.0 upgraded from v0.21.1
- Emissary
v3.3.0 upgraded from v3.0.0
- Metallb
v0.13.7 upgraded from v0.13.5
- Support for packages controller to create target namespaces #601
- (For more EKS Anywhere packages info: v0.13.0
)
Fixed
- Kubernetes version upgrades from 1.23 to 1.24 for Docker clusters (#4266
)
- Added missing docker login when doing authenticated registry pulls
Breaking changes
- Removed support for Kubernetes 1.20
Added
- Add support for Kubernetes 1.24 (CloudStack support to come in future releases)#3491
Fixed
- Fix authenticated registry mirror validations
- Fix capc bug causing orphaned VM’s in slow environments
- Bundle activation problem for package controller
Changed
- Setting minimum wait time for nodes and machinedeployments (#3868, fixes #3822)
Fixed
- Fixed worker node count pointer dereference issue (#3852)
- Fixed eks-anywhere-packages reference in go.mod (#3902)
- Surface dropped error in Cloudstack validations (#3832)
⚠️ Breaking changes
- Certificates signed with SHA-1 are not supported anymore for Registry Mirror. Users with a registry mirror and providing a custom CA cert will need to rotate the certificate served by the registry mirror endpoint before using the new EKS-A version. This is true for both new clusters (
create cluster command) and existing clusters (upgrade cluster command).
- The
--source option was removed from several package commands. Use either --kube-version for registry or --cluster for cluster.
Added
- Add support for EKS Anywhere with provider CloudStack
- Add support to upgrade Bare Metal cluster
- Add support for using Registry Mirror for Bare Metal
- Redhat-based node image support for vSphere, CloudStack and Bare Metal EKS Anywhere clusters
- Allow authenticated image pull using Registry Mirror for Ubuntu on vSphere cluster
- Add option to disable vSphere CSI driver #3148
- Add support for skipping load balancer deployment for Bare Metal so users can use their own load balancers #3608
- Add support to configure aws-iam-authenticator on workload clusters independent of management cluster #2814
- Add EKS Anywhere Packages support for remote management on workload clusters. (For more EKS Anywhere packages info: v0.12.0
)
- Add new EKS Anywhere Packages
- AWS Distro for OpenTelemetry (ADOT)
- Cert Manager
- Cluster Autoscaler
- Metrics Server
Fixed
- Remove special cilium network policy with
policyEnforcementMode set to always due to lack of pod network connectivity for vSphere CSI
- Fixed #3391
#3560
for AWSIamConfig upgrades on EKS Anywhere workload clusters
Added
- Add validate session permission for vsphere
Fixed
- Fix datacenter naming bug for vSphere #3381
- Fix os family validation for vSphere
- Fix controller overwriting secret for vSphere #3404
- Fix unintended rolling upgrades when upgrading from an older EKS-A version for CloudStack
Added
- Add some bundleRef validation
- Enable kube-rbac-proxy on CloudStack cluster controller’s metrics port
Fixed
- Fix issue with fetching EKS-D CRDs/manifests with retries
- Update BundlesRef when building a Spec from file
- Fix worker node upgrade inconsistency in Cloudstack
Added
- Add a preflight check to validate vSphere user’s permissions #2744
Changed
- Make
DiskOffering in CloudStackMachineConfig optional
Fixed
- Fix upgrade failure when flux is enabled #3091
#3093
- Add token-refresher to default images to fix import/download images commands
- Improve retry logic for transient issues with kubectl applies and helm pulls #3167
- Fix issue fetching curated packages images
Added
- Add
--insecure flag to import/download images commands #2878
Breaking Changes
- EKS Anywhere no longer distributes Ubuntu OVAs for use with EKS Anywhere clusters. Building your own Ubuntu-based nodes as described in Building node images
is the only supported way to get that functionality.
Added
- Add support for Kubernetes 1.23 #2159
- Add support for Support Bundle for validating control plane IP with vSphere provider
- Add support for aws-iam-authenticator on Bare Metal
- Curated Packages General Availability
- Added Emissary Ingress Curated Package
Changed
- Install and enable GitOps in the existing cluster with upgrade command
Changed
- Updated EKS Distro versions to latest release
Fixed
- Fixed control plane nodes not upgraded for same kube version #2636
Added
- Added support for EKS Anywhere on bare metal with provider tinkerbell
. EKS Anywhere on bare metal supports complete provisioning cycle, including power on/off and PXE boot for standing up a cluster with the given hardware data.
- Support for node CIDR mask config exposed via the cluster spec. #488
Changed
- Upgraded cilium from 1.9 to 1.10. #1124
- Changes for EKS Anywhere packages v0.10.0
Fixed
- Fix issue using self-signed certificates for registry mirror #1857
Fixed
- Fix issue by avoiding processing Snow images when URI is empty
Added
- Adding support to EKS Anywhere for a generic git provider as the source of truth for GitOps configuration management. #9
- Allow users to configure Cloud Provider and CSI Driver with different credentials. #1730
- Support to install, configure and maintain operational components that are secure and tested by Amazon on EKS Anywhere clusters.#2083
- A new Workshop section has been added to EKS Anywhere documentation.
- Added support for curated packages behind a feature flag #1893
Fixed
- Fix issue specifying proxy configuration for helm template command #2009
Fixed
- Fix issue with upgrading cluster from a previous minor version #1819
Fixed
- Fix issue with downloading artifacts #1753
Added
- SSH keys and Users are now mutable #1208
- OIDC configuration is now mutable #676
- Add support for Cilium’s policy enforcement mode #726
Changed
- Install Cilium networking through Helm instead of static manifest
v0.7.2
- 2022-02-28
Fixed
- Fix issue with downloading artifacts #1327
v0.7.1
- 2022-02-25
Added
- Support for taints in worker node group configurations #189
- Support for taints in control plane configurations #189
- Support for labels in worker node group configuration #486
- Allow removal of worker node groups using the
eksctl anywhere upgrade command #1054
v0.7.0
- 2022-01-27
Added
- Support for
aws-iam-authenticator
as an authentication option in EKS-A clusters #90
- Support for multiple worker node groups in EKS-A clusters #840
- Support for IAM Role for Service Account (IRSA) #601
- New command
upgrade plan cluster lists core component changes affected by upgrade cluster #499
- Support for workload cluster’s control plane and etcd upgrade through GitOps #1007
- Upgrading a Flux managed cluster previously required manual steps. These steps have now been automated.
#759
, #1019
- Cilium CNI will now be upgraded by the
upgrade cluster command #326
Changed
- EKS-A now uses Cluster API (CAPI) v1.0.1 and v1beta1 manifests, upgrading from v0.3.23 and v1alpha3 manifests.
- Kubernetes components and etcd now use TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 as the
configured TLS cipher suite #657
,
#759
- Automated git repository structure changes during Flux component
upgrade workflow #577
v0.6.0 - 2021-10-29
Added
- Support to create and manage workload clusters #94
- Support for upgrading eks-anywhere components #93
, Cluster upgrades
- IMPORTANT: Currently upgrading existing flux managed clusters requires performing a few additional steps
. The fix for upgrading the existing clusters will be published in
0.6.1 release
to improve the upgrade experience.
- k8s CIS compliance #193
- Support bundle improvements #92
- Ability to upgrade control plane nodes before worker nodes #100
- Ability to use your own container registry #98
- Make namespace configurable for anywhere resources #177
Fixed
- Fix ova auto-import issue for multi-datacenter environments #437
- OVA import via EKS-A CLI sometimes fails #254
- Add proxy configuration to etcd nodes for bottlerocket #195
Removed
- overrideClusterSpecFile field in cluster config
v0.5.0
Added