Version 3.0.1

Upcoming Changes

For details on significant changes in future versions, including feature deprecations and removals which may necessitate updates to existing workshops, see Upcoming changes.

New Features

  • Added opinionated installer configuration for Minikube. Will by default install Contour as the ingress controller, but can be disabled if using the nginx ingress controller provided by Minikube.

  • Added opinionated installer configuration for OpenShift. Will use the native OpenShift ingress controller. Will force use of security context constraints (SCC) for cluster security policy engine.

Features Changed

  • When requesting a workshop session using the lookup service, a search for an existing session for that workshop and user is now made against all training portals across all clusters. This is done before filtering out what training portals are accessible to a tenant. This is done so an existing workshop session is found even if labels for clusters or portals had been changed such that the training portal is no longer mapped to the tenant. Such changes to labels may have been made to prevent creation of new workshop sessions against a cluster or training portal when rolling over to a new cluster and waiting for existing workshop sessions on the old cluster to expire. Note that this relies on client side user IDs to be unique across all tenants hosted by the clusters the lookup service monitors. If a single lookup service is being used by multiple distinct custom front end web portals and it is possible the same user ID could be used by more than one front end, it is recommended that the front end incorporate the tenant name as part of the client side user ID passed to the lookup service.

Bugs Fixed

  • When the kind infastructure provider was selected for the opinionated installer, if a CA certificate was provided and installation of educates training platform component was being disabled, the latter was still being installed even though disabled.

  • If a locally managed config file had not been created using the command educates local config edit prior to running educates create-cluster, it would fail as the local config directory wasn’t being created automatically.

  • Using the custom provider when the educates component was enabled, would fail when using CLI to create or update a cluster as image references were not being replaced with references to where released image artifacts were stored.

  • When an ingress secret was not being supplied, but clusterIngress.protocol was being forcibly set to https due to HTTPS connections being terminated by a front end router, annotations were wrongly being added to ingresses to force redirection from http to https. This would result in a redirect loop. Annotations should not have been added and redirection from http to https should instead have been done by the front end router.