CHANGELOG-2.0.md

Changelog 2.0

[2.0.1] YYYY-MM-DD

Added

• #2932 - Support epicli upgrade for RHEL/AlmaLinux 8
• #3057 - Additional AWS authentication option
• #3101 - Add support for ARM architecture for AlmaLinux 8
• #3105 - Add manifest file parsing
• #3131 - Optimize Grafana dashboards downloading
• #3116 - Optimize files downloading
• #3106 - Add image-registry configuration reading
• #3140 - Allow to disable OpenSearch audit logs
• #3218 - Add support for original output coloring
• #3079 - OpenSearch improvement - add dedicated user for Filebeat
• #3207 - Add filtering mechanism for the sensitive data

Fixed

• #3153 - AlmaLinux 8.5 installation fails resolving dependencies
• #3164 - Specify version and allow containerd.io package downgrade in haproxy_runc role
• #3179 - terraform fails when use_network_security_groups is set to false
• #3165 - download-requirements.py may fail due to expired certificate
• #3189 - Fix configuration/feature-mapping enabling
• #3152 - Use a stable tag for the quay.io/ceph/ceph:v16.2.7 image
• #3209 - [Ubuntu] download-requirements.py ignores package version when resolving dependencies
• #3231 - epicli may fail on "Download image haproxy-2.2.2-alpine.tar" task
• #3210 - [Ubuntu] download-requirements.py downloads redundant package dependencies
• #3190 - Enable configuration of kubelet enable-controller-attach-detach argument via input manifest

Updated

• #3080 - update Filebeat to the latest compatible version with OpenSearch
• #2982 - Using AKS and EKS Terraform configuration directly with Epiphany.
• #2870 - OpenDistro for ElasticSearch replaced by OpenSearch
• #3163 - Upgrade Python dependencies
• #3097 - Split available_roles and roles_mapping into separate yaml documents
• #3229 - Update crane to v0.11.0

Deprecated

• Support for Modules: Azure Basic Infrastructure (AzBI) module Azure AKS (AzKS) module AWS Basic Infrastructure (AwsBI) module AWS EKS (AwsKS) module

Breaking changes

• Schema configuration/feature-mapping changed. The document was splitted into two separate docs configuration/features and configuration/feature-mappings.

• AWS credentials configuration parameters are renamed from specification.cloud.credentials.key and specification.cloud.credentials.secret to specification.cloud.credentials.access_key_id and specification.cloud.credentials.secret_access_key.

[2.0.0] 2022-05-09

Added

• #1332 - Support for RHEL 8.4 and AlmaLinux 8.4
• #959 - Add usage of use_network_security_groups to disable NSG on AWS
• #2701 - Epicli prepare - generate files in separate directory
• #2812 - Extend K8s config validation
• #2950 - CLI refactor to make it more consistent
• #2844 - Refactor K8s upgrade task in order to simplify its flow
• #2985 - Make RabbitMQ Plugins configurable
• #2974 - Refactor Apply command
• #2976 - Allow for custom Terraform scripts
• #2716 - Change container runtime to containerd
• #805 - Refactor download-requirements script
• #2858 - Make Ruby spec tests code compliant with rubocop lint rules
• #2975 - Copy only required files
• #2991 - Add automatic backup creation for download requirements
• #2448 - Passwordless SSH communication for postgres user between DB nodes
• #2821 - Node Exporter preflight checks
• #2996 - Introduce the new configuration field to change a component name
• #2888 - Define additional disks with defined sizes to VMs for Azure
• #2812 - Extend K8s config validation
• #2890 - Rook/Ceph Kubernetes Native Storage for Azure Kubernetes non-AKS cluster
• #1452 - Create Kubernetes Native Storage Epiphany configuration for Azure provider
• #2891 - Rook/Ceph Kubernetes Native Storage for AWS non-EKS
• #2887 - Create Kubernetes Native Storage Epiphany configuration for AWS provider
• #2734 - Document upgrade Red Hat / CentOS 7 to v 8.x

Fixed

• #2653 - Epicli is failing in air-gapped infra mode
• #1569 - Azure unmanaged disks not supported by Epiphany but there is misleading setting in the default configuration
• #2832 - Make the DoD checklist clear
• #2853 - Change autoscaling_group approach in AWS provider in favor of plain VM creation.
• #2669 - Restarting the installation process can cause certificate problems if K8s was not fully configured
• #2944 - Refactor InitEngine class to be agnostic to changes in ApplyEngine and UpgradeEngine
• #2945 - epicli apply sleeps 10 seconds after creating inventory
• #2968 - epicli init should generate specification.cloud.subscription_name for minimal cluster config
• #2940 - firewalld.service unit could not be found on host however ansible_facts sees it as defined
• #2979 - Restore the possibility of choosing the availability zone in AWS
• #2984 - Validation blocks overwriting of destination_address_prefix in NSG rules, which is 0.0.0.0/0 by default
• #2966 - epicli init --full does not generate configuration for OpenDistro
• #2942 - rsync command fails trying to copy artifacts
• #2930 - Backup/recovery commands fail when default configuration for backup attached to cluster-config.yml
• #2989 - Task Remove swap from /etc/fstab does not remove swap entry from file
• #2907 - Backup/recovery commands fail when executed directly after upgrade
• #3025 - Running yum commands may hang waiting for user input
• #2728 - PostgreSQL's configuration files located outside the data directory are not copied by repmgr
• #3029 - [RHEL] Single machine upgrade fails on preflight check: 'dict object' has no attribute 'size_available'
• #2997 - Doubled HAProxy entries in prometheus.yml after upgrade
• #2803 - Refactor: rename 'kafka_var' setting
• #2995 - Update expired RHUI client certificate before installing any RHEL packages
• #3049 - HAProxy upgrade fails on re-run trying to remove haproxy_exporter
• #3006 - install containerd.io=1.4.12-* failed, when upgrade from v1.3.0 to 2.0.0dev
• #3065 - Flag delete_os_disk_on_termination has no effect when removing cluster

Updated

• #2828 - K8s improvements
  • Re-generate apiserver certificates only by purpose
  • Do not ignore preflight errors in kubeadm join
  • Update documentation about control plane certificates renewal
• #2825 - Upgrade Terraform and providers
  • Terraform 0.12.6 to 1.1.3 (#2706)
  • Azurerm provider 1.38.0 to 2.91.0
  • AWS provider 2.26 to 3.71.0
  • Upgraded Azure-cli 2.29 to 2.32
• #2847 - Upgrade Ansible to 5.2.0
  • Ansible 2.10.15 to 5.2.0
  • Python 3.7 to 3.10
• #2871 - Upgrade Kafka to 2.8.1
• #3090 - Update cloud OS images to the latest

Removed

• Support for RHEL 7
• Support for CentOS
• #2834 - Removal of Hashicorp Vault component
• #2833 - Removal of Logstash component
• #2836 - Removal of Istio component
• #2837 - Removal of Apache Ignite component
• #2927 - Review Epiphany tools (remove outdated)
• #2864 - Remove unnecessary design docs

Deprecated

Breaking changes

• RHEL 7 and CentOS are not supported. For RHEL, release 8.4 is required and upgrade from version 7 is not handled by Epiphany. Support for CentOS has been replaced by AlmaLinux (migration is not handled by Epiphany).
• Support for ARM architecture has been removed (together with CentOS), planned to be added for AlmaLinux in the next release.
• Upgrade of Terraform components in issue #2825 and #2853 will make running re-apply with infrastructure break on existing 1.x clusters. The advice is to deploy a new cluster and migrate data. If needed a manual upgrade path is described here.
• Kubernetes container runtime changed. Dockershim and Docker are no longer on Kubernetes hosts.
• Filebeat docker input replaced by container input. New field provided for Filebeat as system service installation: container.id. Field kubernetes.container.name is no longer valid.

Known issues

• epicli upgrade is not supported yet for RHEL or AlmaLinux