1.8 (Clio)

Changelog for 1.8 (Codename: Clio)

Starting with 1.8, Director now provides a GUI wizard to install latest OpenEBS in a connected cluster.
Users can connect any Kubernetes cluster (above v1.12) to Director and install OpenEBS on their cluster using the new wizard.
Two modes of installation are supported - Basic and Advanced.
In Basic mode, OpenEBS will be installed with a default configuration.
In Advanced mode, users can specify a custom namespace, image registry and data directory, specify a set of nodes where OpenEBS control-plane component pods will be scheduled, specify a set of nodes where OpenEBS data-plane component pods will be scheduled and specify include/exclude device filter configuration for Node Device Manager (NDM).
Only the latest version of OpenEBS eg. 1.8 is supported for installation.
A cluster cannot have more than one instance of OpenEBS running.
A new Data Agility Operator (DAO) component named 'openebs-upgrade' handles the installation of OpenEBS in the connected cluster. This operator is a Kubernetes custom controller based on metacontroller add-on.
For more information on OpenEBS install support in Director, visit here.

Director (prior to 1.8) set up alerting for OpenEBS and other components like volumes, pools, pods, nodes with pre-defined rules which evaluate metrics, states and logs of those components.
In 1.8, these rules are now visible to the user.
User can now view all the rules which lead to the generation of alerts.
In addition, a user has granular control like enabling or disabling of the rules and also configuring the threshold for some of the rules.
The rules are categorized into metric-based and log-based.
Rules can be enabled, disabled or modified at a cluster level.
For more information on Alerts and Alert Rules in Director, visit here.

Director 1.8 has an enhanced UI for displaying the alerts.
Alerts are now scoped at cluster-level. The 'Notifications' link under 'Project' scope in UI side-bar has been temporarily removed as it was leading to duplicate content.
Alerts are categorized under Acknowledged and Unacknowledged.
In addition to the support for timestamp-based filtering, alerts can now be filtered based on the component types like pools and volumes.
From an alert, the user can now navigate to the rule that has triggered the alert.
For more information on Alerts and Alert Rules in Director, visit here.

Director 1.8 has an enhanced v2 schema for the alerts.
The new schema has provision to store the 'Viewed' property, 'Component Type' and 'Component Identification' in addition to the older schema elements.
An upgrade of Director will result in seamless migration of all v1 alerts to v2.

Director 1.8 supports upgrade of OpenEBS pools and volumes to their version 1.8.
For more information related to upgrade, visit here.

Director OnPrem can now be deployed without cloning the Director Repository using Packaged-helm method. For more information, visit here.

Fixed an issue which was not allowing a user to disconnect the Director cluster from UI.
Fixed an issue due to which a user, after changing his email, was not able to login with the changed email. However, this does not apply to the Director administrator, he still needs to login with the 'administrator' username.
Fixed an issue which resulted in broken monitoring graphs for the Director cluster.
Fixed an issue which resulted in incorrect node details for cStor volume replicas of statefulset applications.
Fixed an issue which resulted in Cassandra pod going to CrashLoopBackoff state when Director is deployed on Konvoy-based Kubernetes clusters.

A user, who is a part of multiple projects, will not be able to open monitoring graphs and logs of different projects in different browser tabs.
Auto-refresh of the Alerts page over-rides the user-selected filters on the page.
Currently, Minio application monitoring only works for Minio with image minio/minio:RELEASE.2019-09-18T21-55-05Z and earlier. The images later than minio/minio:RELEASE.2019-09-18T21-55-05Z have an in-built authentication that needs to be supported by Director.
Sometimes, monitoring graphs display error if Director consul or ingester pod restarts. The workaround is to manually restart the consul pod and all the ingester, querier and distributor pods.