Chapter 2. Known issues

This section lists known issues with Red Hat CodeReady Workspaces 2.1. Where available, workaround suggestions are provided.

2.1. Workspaces based on java-eap-maved created in CodeReady Workspaces 2.0 fail to start after the update to 2.1.1 on OpenShift 3.11

Users with the cluster-admin role who use OpenShift 3.11 may experience the 2.0 version java-eap-maved-based workspaces with TLS and OpenShift OAuth support fail to start after the product server is updated to version 2.1.1. Users are then not permitted to access the old Persistent Volume.

To work around this issue, create a new 2.1.1 workspace using the same devfile that was used to create the workspace in the previous version.

2.2. Openshift Connector plug-in requires manual connecting to the target cluster

By default, the Openshift Connector plug-in logs into the cluster as inClusterUser, which may not have the manage project permission. This causes an error message to be displayed when a new project is being created using Openshift Application Explorer:

Failed to create Project with error 'Error: Command failed: "/tmp/vscode-unpacked/redhat.vscode-openshift -connector.latest.qvkozqtkba.openshift-connector-0.1.4-523.vsix/extension/out/tools/linux/odo" project create test-project ✗ projectrequests.project.openshift.io is forbidden

To work around this issue, log out from the local cluster and relog in to OpenShift cluster using the OpenShift user’s credentials.

2.3. Missing debug configuration for the Node JS Config Map devfile workspaces

Absence of the debug configuration, situated in the debug panel of a created workspace, when the Node JS Config Map devfile is used for a new workspace creation.

2.4. CodeReady Workspaces is failing to shut down after executing crwctl server:stop

The dedicated crwctl command crwctl server:stop is unable to shut down the CodeReady Workspaces server and instead fails with a timeout and displays the following error message:

›   Error: E_SHUTDOWN_CHE_SERVER_FAIL - Failed to shutdown CodeReady Workspaces server. E_CHE_API_NO_RESPONSE - Endpoint: http://codeready-ndp-test.apps.crw.codereadyqe.com/api/system/stop?shutdown=true - Error message: timeout of
 ›   3000ms exceeded

To work around the issue, execute crwctl server:stop again.

2.5. The crwctl workspace:inject command does not work in workspaces with OpenShift OAuth support

The crwctl workspace:inject command causes the error during the workspace Pod localization in workspaces created with OpenShift OAuth support.

$ crwctl workspace:inject -n codeready-tls-oauth -k
  ✔ Verify if namespace codeready-tls-oauth exists
  ✖ Verify if the workspaces is running
    → No workspace pod is found
    Injecting configurations
 ›   Error: No workspace pod is found

To work around the issue, use oc login command inside the affected container instead.

2.6. Openshift Projects remain present after a namespace deletion CodeReady Workspaces 2.1.1

When a workspace has been created in a dedicated namespace that is later entirely deleted, the corresponding OpenShift project needs to be removed manually to complete the deletion process.

To work around the issue: delete the remaining empty project from the OpenShift console:

$ oc delete project <projectname>

2.7. The crwctl server:delete uninstallation command does not remove the OpenShift project

After using the crwctl server:delete command, the OpenShift project that used to host the CodeReady Workspaces instance remains. This makes it impossible to install a new CodeReady Workspaces instance into the default namespace, which still exists. To uninstall CodeReady Workspaces completely, manually remove the namespace.

To work around the issue:

  1. Stop the Red Hat CodeReady Workspaces Server:

    $ crwctl server:stop
  2. Obtain the name of the CodeReady Workspaces namespace:

    $ oc get checluster --all-namespaces -o=jsonpath="{.items[*].metadata.namespace}"
  3. Remove CodeReady Workspaces from the cluster:

    $ crwctl server:delete -n <namespace>

    This removes all CodeReady Workspaces installations from the cluster.

  4. Delete the checluster object and the codeready-workspaces resource:

    $ oc delete checluster codeready-workspaces --namespace=<openshift_namespace>

    <openshift_namespace> is the name of the OpenShift project where CodeReady Workspaces is deployed.

  5. Delete the OpenShift namespace:

    $ oc delete project <openshift_namespace>

2.8. Uninstallation command crwctl server:delete might make OpenShift instance unusable

Creation of the CodeReady Workspaces cluster, codeready-workspaces, in a namespace can be affected by previous use of the crwctl server:delete uninstallation command.

To work around the issue:

  1. Patch the Custom Resource Definition:

    $ oc patch customresourcedefinition/checlusters.org.eclipse.che -p \
    '{ "metadata": { "finalizers": null }}' --type merge

2.9. An embedded application of the "Java EAP Maven" stack tends to fail at launch in the debug mode

An embedded application of the Java EAP Maven stack in the debug mode tends to fail. The dialog window with application URL is already displayed, but the application, according to the terminal output, is still starting. The use of the URL link displayed leads to an error.

2.10. Entering a workspace fails after restarting it

Attempting to restart a workspace and re-enter it fails, and an error message is displayed instead. To work around this issue, restart the workspace again.

2.11. Workspace Cap and Workspace RAM Cap organization restrictions do not work

The Workspace Cap and Workspace RAM cap functions, which control the maximum number of workspaces for an organization and the maximum RAM that organization workspaces can use, currently do not work.

2.12. The terminal tab for a workspace in some cases does not open

Devfiles contain a set of predefined commands that can be executed in workspaces started using devfiles from the devfile registry. However, when a command defined by a devfile is executed from the workspace, the terminal, in which the commands normally run, does not open.

The work around this issue, do not open the same workspace link in two different browsers.

2.13. Workspaces are not stopped by the idling timeout

Due to CodeReady Workspaces and OpenShift OAuth integration, workspaces are not stopped by the idling timeout, when the workspace is located in a user’s Pod.

To work around this issue, disable the stopping of idling workspace by timeout feature for the workspaces with OpenShift OAuth integration:

$ oc patch checluster/eclipse-che --patch "{\"spec\":{\"server\":{\"customCheProperties\": {\"CHE_LIMITS_WORKSPACE_IDLE_TIMEOUT\": \"-1\"}}}}" --type=merge -n che

2.14. Error highlighting and code completion do not work in a Go devfile

To workaround this issue, update the Go language server plug-in to the latest version.

2.15. Debugging utility does not work correctly on its first run in the Go devfile workspace

In workspaces based on the Go devfile, an error notification is displayed during the start of the Debug current file configuration.

To work around this issue, execute the predefined Run current file command first, then repeat the debugging procedure.