Release Notes

Management adds a Nitrokey option for an initialized Nitrokey HSM2, with a token serial and the appliance node to which the USB device is attached. All CARA replicas use the token on the selected node. Loss of that node or token interrupts HSM access, including on a three-node appliance.

Both single-node and clustered installations require a maintenance window for this update. There is no direct 1.0.0-to-1.0.2 update bundle.

Update Management to 1.0.2, then complete the appliance update before making other GitOps changes. Until the update converts the HSM values, Management cannot read the old ERS settings and Operations rejects ordinary GitOps publication. The update’s own publication runs after conversion.

Follow the update procedure and preserve the verified source recovery set. The update verifies the original 1.0.1 backup without rewriting it. Source backups remain recoverable with 1.0.1 release content; target backups require 1.0.2 release content. Keep both recovery sets with their matching release bundles.

The update loads changed offline content, replaces vendor declarations while retaining installation settings, and waits for the published configuration and workloads to converge. It retains explicit Retry for interrupted scripts or publication failures.

The following table lists selected version changes. The versioned release inputs record the complete image and chart selections with their checksums and digests.

The ERS chart remains at 2.1.0, CARA at 3.6.0, CLM at 6.15.0 and Keycloak at 26.7.3.

The update bundle supplies the changed offline images and charts, including the matching Talos installers. Its numbered scripts update the platform, publish the target configuration, and verify component readiness before recording appliance version 1.0.1. The update preserves installation settings and selects the installer for the existing agentless, QEMU-agent or VMware-agent profile. Secure Boot configurations are outside this update’s supported installer profiles.

Follow the update procedure with Management 1.0.0 or a version explicitly qualified for this appliance. Keep the exact source recovery set outside Management and acknowledge it before applying the update. An interrupted script or failed publication requires explicit Retry. After completion, retain the new target backup and verify ERS access.

Single-node installations require maintenance. The published bundle declares that maintenance is not required for clustered installations.

Use the 1.0.1 release bundle for a fresh installation or recovery of a matching backup. A release bundle cannot replace the 1.0.0-to-1.0.1 update bundle.

Deploy Management using the Management ISO or Docker Compose, then follow appliance installation and verification.

For Management ISO deployments, SSH is enabled with factory credentials for initial setup convenience. You must restrict SSH before first boot, configure it according to your organization’s policy, rotate the factory host credentials and verify persistence after reboot before production use. Follow Required ISO host SSH setup.

Retain the downloaded artifact bundle outside Management. Set up backups and protected external retention, record the deployed component identities and verify application access after ERS setup.

Management access is temporary. Keep the artifact bundle available for reconnection after access expiry or Management restart. Management loss and appliance loss use different recovery paths.

There is no upgrade path from an earlier released Virtual Appliance version because 1.0.0 is the first release. Migration from development media or state is not supported.

Use the supplied 1.0.0 release material to identify compatible Management media and embedded component versions. See components and version records. Full recovery requires the release bundle that matches the backed-up appliance release, together with its matching artifacts and compatible Management.