Releases: aws-greengrass/aws-greengrass-nucleus
Releases · aws-greengrass/aws-greengrass-nucleus
2.11.1
Bug fixes and improvements
- Fixes an issue where the nucleus may not start if a bootstrap task fails and the deployment metadata file is corrupted
- Fixes an issue where on-demand Lambda components aren't reported in deployment status updates
- Adds support for duplicate authorization policy IDs
2.11.0
2.10.3
2.10.2
Bug fixes and improvements
- Allows case insensitive parsing of component lifecycle
- Fixes an issue where environment PATH variable was not getting recreated correctly.
- Fix proxy URI encoding for components including stream manager for usernames with special characters.
2.10.1
Bug fixes and improvements
- Revert SDK CRT to fix launch failure on some armv8 processors
- Check configured input stream option before close
2.10.0
New features
- Adds support for MQTT5.
- Adds a mechanism for loading plugin components quickly without scanning.
- Enables Greengrass to save disk space by deleting unused Docker images.
- Allows access to private Amazon ECR repositories in regions other than the AWS IoT Greengrass core region.
- Adds an optional configuration value to define where the IPC socket file should be created on Linux.
Bug fixes and improvements
- Fixes an issue where rollback leaves certain configuration values in place from a deployment.
- Fixes an issue where the Greengrass nucleus validates for an AWS domain sequence in custom non-AWS credentials and data endpoints.
- Updates multi-group dependency resolution to re-resolve all group dependencies via AWS Cloud negotiation, instead of locking to the active version. This update also removes the deployment error code INSTALLED_COMPONENT_NOT_FOUND.
- Updates the Greengrass nucleus to skip downloading Docker images when they already exist locally.
- Updates the Greengrass nucleus to restart a component install step before timeout expires.
- Additional minor fixes and improvements.
2.9.6
Bug fixes and improvements
This update fixes an issue where a Greengrass deployment fails with the error LAUNCH_DIRECTORY_CORRUPTED and a subsequent device reboot would fail to start Greengrass. This error may occur when you move the Greengrass device between multiple thing groups with deployments that require Greengrass to restart.
2.9.5
New features
- Adds support for Greengrass nucleus software signature verification.
Bug fixes and improvements
- Fixes an issue where a deployment fails when the local recipe metadata region doesn't match the Greengrass nucleus launch region. The Greengrass nucleus now renegotiates with the cloud when this happens.
- Fixes an issue where the MQTT message spooler fills up and never removes messages.
- Additional minor fixes and improvements.
2.9.4
Bug fixes and improvements
- Checks for a null message before it drops QOS 0 messages.
- Truncates job status detail values if they exceed the 1024 character limit.
- Updates the bootstrap script for Windows to correctly read the Greengrass root path if that path includes spaces.
- Updates subscribing to AWS IoT Core so that it drops client messages if the subscription response wasn't sent.
- Ensures that the nucleus loads its configuration from backup files when the main configuration file is corrupt or missing.
2.9.3
Bug fixes and improvements
- Ensure MQTT client IDs are not duplicated
- Add more robust file-reading/writing to avoid and recover from corruption
- Retry Docker image pull on specific network-related errors
- Enable the noProxyAddresses option for MQTT connection