v1.3.2

What's Changed

• Update README for 1.3.z by @lzichong in #654
• CAPI-1.7.4: Move core-CAPI dependencies to v1.7.4. by @arunmk in #659
• Add TKG 2.5.2 cluster template files by @lzichong in #660
• 650: Fix issue with CSI manifests for v1.6.0 by @arunmk in #663
• Vendor update on removing remove eng.vmware.com by @sakthisunda in #662
• Remove v1alpha4 from CRDs by @lzichong in #664

Full Changelog: v1.3.1...v1.3.2

v1.3.1

What's Changed

• [CAFV-449][1.3.z] update CLUSTERCTL.md with updated versions by @rliang88 in #614
• [CAFV-460][1.3.z]update CLUSTERCTL.md with TKG catalog management table by @rliang88 in #624
• [CAFV-466][1.3.z] bring CLUSTERCTL.md and CRS.md up to date by @rliang88 in #633
• [CAFV-464][1.3.z]Add suggested CRS labels to crs.yaml files by @rliang88 in #638
• [CAFV-464][1.3.z] Remove v1.20.8 crs and cluster template files by @rliang88 in #641
• Consume updated common-core to fix issues around ip spaces by @rocknes in #651

Full Changelog: v1.3.0...v1.3.1

v1.3.0

Features:

• Upgrade CAPI from v1.4.z to v1.5.4 (#596)
• Consume latest common-core from CPI 1.6.0 to enable IP space support (#599)
• TKG 2.5 templates for clusterctl (#607)
• Deprecate VCD API v36.x in CAPVCD (#580)

Changes in Behavior:

• Bump golang.org/x/crypto from 0.14.0 to 0.17.0 (#570)

v1.2.0

Features:

• Add TKG 2.3.1 and TKG 2.2.0 support
• Added clusterctl templates for Kubernetes v1.27, v1.26 versions [#546]

Changes in Behavior:

• Bump golang.org/x/net from 0.8.0 to 0.17.0 [#532]

Bug Fixes:

• Use machine.Name so that vm.VM.name will not cause panics [#564]

v1.1.1

Features:
None. This is a patch release

Changes in Behavior:

1. VMs created by CAPVCD will begin to be created with 15 character passwords for the root user (PR #517). Earlier they were created with 8 characters.

Bug Fixes:

1. CAPVCD earlier installed nvidia-container-runtime package and edits containerd runtime to point to nvidia-container-runtime during the provisioning of vGPU node. This customization interfered with Nvidia GPU Operator installation. The fix (PR #525) removes the customization in favor of GPU Operator installation. Users who wish to manually install drivers on vGPU node are required to do the customization themselves - (a) install the nvidia-container-runtime package and (b) update the containerd configuration to point to the NVIDIA runtime.
2. There was an issue in the cloud-init script where containerd service was restarted and then a few ctr commands were executed before the containerd service was up. This led to a race condition in some rare cases where containerd would take some time to start up. This has been fixed.
3. The enable-hostpath-provisioner parameter has been removed from all manifest files since they were never consumed by CAPVCD
4. PodSecurityPolicy related entries in manifests have been removed since they have been removed from Kubernetes v1.25 onwards.

v1.1.0

Features:

1. TKG 2.1.1, TKG 2.2.0 Support
2. Updated CAPI version to v1.4.0 (PR #445)
3. CAPVCD API version changed to v1beta2 (PR #374)
4. Clusterclass Support (PR #357)
5. Golang 1.19 Support (PR #440)
6. Customized naming convention for cluster VMs (PR #369)
7. Qualified CSI 1.4.0, CPI 1.4.0
8. CAPVCD can support OVDC name changes post cluster creation (PR #442)
9. Support multiple NIC for machines (PR #254)
10. Added clusterctl templates for Kubernetes v1.25, v1.24 versions (PR #476)

Misc/Bug Fixes:

1. CAPVCD Upgrade Documentation
2. Reduce image size, and move to Golang static binary and scratch image (PR #446)
3. Added test automation for workload cluster CRUD operations (PR #475)
4. Qualified MachineHealthCheck and Cluster Autoscaler
5. Please refer to the TKG documentation for the Antrea installation as CRS templates are not supported for TKG 2.1.1, TKG 2.2.0.
6. CAPVCD now closes idle http connections when reconciliation ends (PR #468)

v1.0.2

Bug Fixes from v1.0.1

1. Support for TKG 1.6.1 cluster attachment to Tanzu Mission Control (PR #447)

Others

1. Bump github.com/docker/distribution 2.8.1 to 2.8.2 (PR #450)

v1.0.1

Features

1. Added Support for TKG 1.6.1

Bug Fixes from v1.0.0

1. Fix multiple OVDC with same network names by adding OVDC checks when getting OVDC network (PR #399)
2. Prevent CAPVCD from crashing when using IP Spaces (PR #372)
3. Ensure CAPVCD controller accepts RDE 1.1.0 for cluster creation (PR #402)
4. Fix for nil pointer exception on fetching OVDC network (PR #429)

v1.0.0

Features:

1. Bump CAPI version to v1.1.3
2. Multi-version API support; CAPVCD API version bumped up to v1beta1
3. Custom Resource Set definitions to install add-ons like CPI, CNI, CSI
4. NVIDIA GPU-enabled clusters
5. Live visibility into CAPVCD clusters from VCD UI
6. Proxy support for the workload clusters
7. User ability to specify node-pools with distinct disk size, sizing policy and placement policy for both control plane nodes and worker nodes.
8. User ability to specify Control plane endpoint. Private IP support for control-plane endpoint using this feature.
9. clusterctl support with some manual configuration.
10. User ability to specify VCD credentials via Secrets.
11. Support upgrades from CAPVCD 0.5.x clusters to CAPVCD 1.0.0 clusters
12. Support user-specified VIP subnet propagation to CPI
13. CAPVCD start-up with zero configuration
14. Control-plane Loadbalancer creation with no DNAT rule
15. Metering statistics collection
16. storageProfile selection for VCDMachine objects

v0.5.1

Bug Fixes from v0.5.0

• Fixed system/administrator not being able to create clusters due to username being passed in full during authentication (PR #117)
• Fixed authentication issue with refresh tokens as a system/administrator user (PR #119)
• Fixed controllers using CAPVCD org to place workload clusters instead of the org user belongs to (PR #88)
• Restart nodes in cloud-init if files are absent (PR #53)
• Ensure that Antrea is downloaded from harbor repo (PR #75)
• Ensure that VCD clients are long-lived