General known issues

This page contains information regarding general known issues and their proposed resolution.

Issues described in this document are general and can occur irrespectively of using development or release version.

Issues that are specific to process of installing certain release and migrating between versions are described in separate document respective to version. All documents related to specific version issues are located in this folder.

Known issues of release 0.19

Known issues of release 0.20

Description

Issues are separated depending on implementation they belong to.

Current implementations:

Google Cloud - Terraform-based.
Baremetal Implementation - Helm-based.

Google Cloud Implementation Known Issues

Failing `terraform apply` command

During installation of Google Cloud implementation of OSDU using Terraform module or simple installation of named Terraform module main command responsible for installation of OSDU terraform apply can fail due to unstable internet connection, internal API problems and other unpredictable issues.

Default resolution would be to look at previous command output indicating possible problems, solve them or simply reapply the same command:

terraform apply

Inability to change type of domain installation (custom or default Endpoints)

Currently Google Cloud implementation supports both custom domain (set using "domain" variable during terraform installation) and default domain from Endpoints service.

Currently it's not possible to switch between these options after installation. Change from one type of domain to another requires either re-install of Google Cloud Implementation or some manual actions.

Baremetal Implementation Known Issues

Some services are failing after installation

Baremetal implementation uses Helm to install all needed infrastucture components and OSDU services. It installs all elements at once, irrespective of internal dependencies (e.g. many services need provisioned entitlements service to work). Due to that fact, it may take significant time (up to 30 minutes) for environment to stabilize and for all services to wait until their dependencies are up and ready.

As bootstrap of infrastructure elements and services is idempotent, numerous restarts of services do not influence or corrupt data.

Proposed resolution would be to wait until all installed deployments have status Ready. To list all installed deployments and see their status, the following command can be used:

kubectl get deployments