# v1.7.0

## Release Notes

### New Features & Enhancements

#### PVH support

Add support for allocation of 1 GB and 2 MB pages to improve performance.

#### Installer

Installer now checks currently mounted host filesystems to dynamically build the correct root filesystem drivers into the initrd for initial boot.

#### Kubernetes

Basic support for Flannel CNI.

#### Miscellaneous

- Report Edera version number as part of `CONTAINER-RUNTIME` field with `kubectl get nodes -o wide`.

## Bug Fixes

### Kubernetes integration

- Fixed issue where container termination messages were not always compliant with other CRI runtimes.
- Fixed issue where pod log flushing behavior was not always correct.
- Fixed issue where containers in the same pod did not share the same IPC namespace by default.
- Fixed issue where Edera pods in a Kubernetes dual-stack cluster would not report all CNI-assigned IP addresses to Kubernetes.
- Fixed issue where `imagePullPolicy: Always` might not fully bust the cache.
- Fixed issue where Flannel and other CNI plugins did not work, due to incorrect subnet mask mirroring.
- Ensure `protect-cri` systemd service starts before kubelet.
- Fixed issue where Edera-managed pod `stderr` logs would not propagate correctly via `kubectl logs`.

### Core

- Fixed issue where PVH shadow pool size was incorrect, leading to PVH zones not starting.
- Fixed issues with vCPU bring-up order under PVH.
- Fixed issue where very long symlink targets in OCI images might be incorrectly truncated due to lack of extended PAX attribute propagation.
- Fixed issue where `mksquashfs` process might not cleanly terminate in all cases.
- Fixed issue where a cached image manifest might still prompt an unnecessary remote lookup.
- Ensure `protect-daemon` systemd service restart doesn’t unnecessarily restart `protect-storage` service.

### Installer fixes

- Fixed issue where air-gapped installer was missing some newly added components.
- Fixed issue where `dracut` initrd generation might fail silently if `/boot` is full.

### Falco Plugin

- pid 1 is now excluded by default from kernel syscall event monitoring.
- Zone kernel 6.16 or newer is no longer required for Falco event streaming—any Edera zone kernel, including the default one, is supported for Falco event streaming.

## Known issues

- PVH support is still in beta:
  - The static resource policy is nonfunctional for PVH zones.
  - K8S manifest resource allocation is nonfunctional under PVH.

## Upgrade notes

There are no known breaking changes in this release from the previous release `v1.6.0`.

Last updated on 2026-03-10