Chapter 7. Resolved issues

The following notable issues are resolved in Red Hat OpenShift AI.

RHOAIENG-15033 - Model registry instances do not restart or update after upgrading OpenShift AI

Previously, when you upgraded OpenShift AI, existing instances of the model registry component were not updated, which caused the instance pods to use older images than the ones referenced by the operator pod. This issue is now resolved.

RHOAIENG-15008 - Error when creating a bias metric from the CLI without a request name

Previously, the user interface sometimes displayed an error message when you viewed bias metrics if the requestName parameter was not set. If you used the user interface to view bias metrics, but wanted to configure them through the CLI, you had to specify a requestName parameter within your payload. This issue is now resolved.

RHOAIENG-14986 - Incorrect package path causes copy_demo_nbs to fail

Previously, the copy_demo_nbs() function of the CodeFlare SDK failed because of an incorrect path to the SDK package. Running this function resulted in a FileNotFound error. This issue is now resolved.

RHOAIENG-14552 - Workbench or notebook OAuth proxy fails with FIPS on OCP 4.16

Previously, when using OpenShift 4.16 or newer in a FIPS-enabled cluster, connecting to a running workbench failed because the connection between the internal component oauth-proxy and the OpenShift ingress failed with a TLS handshake error. When opening a workbench, the browser showed an "Application is not available" screen without any additional diagnostics. This issue is now resolved.

RHOAIENG-14095 - The dashboard is temporarily unavailable after the installing OpenShift AI Operator

Previously, after you installed the OpenShift AI Operator, the OpenShift AI dashboard was unavailable for approximately three minutes. As a result, a Cannot read properties of undefined page sometimes appeared. This issue is now resolved.

RHOAIENG-13633 - Cannot set a serving platform for a project without first deploying a model from outside of the model registry

Previously, you could not set a serving platform for a project without first deploying a model from outside of the model registry. You could not deploy a model from a model registry to a project unless the project already had single-model or multi-model serving selected. The only way to select single-model or multi-model serving from the OpenShift AI UI was to first deploy a model or model server from outside the registry. This issue is now resolved.

RHOAIENG-545 - Cannot specify a generic default node runtime image in JupyterLab pipeline editor

Previously, when you edited an Elyra pipeline in the JupyterLab IDE pipeline editor, and you clicked the PIPELINE PROPERTIES tab, and scrolled to the Generic Node Defaults section and edited the Runtime Image field, your changes were not saved. This issue is now resolved.

RHOAIENG-14571 - Data Science Pipelines API Server unreachable in managed IBM Cloud OpenShift OpenShift AI installation

Previously, when configuring a data science pipeline server, communication errors that prevented successful interaction with the pipeline server occurred. This issue is now resolved.

RHOAIENG-14195 - Ray cluster creation fails when deprecated head_memory parameter is used

Previously, if you included the deprecated head_memory parameter in your Ray cluster configuration, the Ray cluster creation failed. This issue is now resolved.

RHOAIENG-11895 - Unable to clone a GitHub repo in JupyterLab when configuring a custom CA bundle using |-

Previously, if you configured a custom Certificate Authority (CA) bundle in the DSCInitialization (DSCI) object using |-, cloning a repo from JupyterLab failed. This issue is now resolved.

RHOAIENG-1132 (previously documented as RHODS-6383) - An ImagePullBackOff error message is not displayed when required during the workbench creation process

Previously, pods experienced issues pulling container images from the container registry. When an error occurred, the relevant pod entered into an ImagePullBackOff state. During the workbench creation process, if an ImagePullBackOff error occurred, an appropriate message was not displayed. This issue is now resolved.

RHOAIENG-13327 - Importer component (dsl.importer) prevents pipelines from running

Pipelines could not run when using the data science pipelines importer component, dsl.importer. This issue is now resolved.

RHOAIENG-14652 - kfp-client unable to connect to the pipeline server on OCP 4.16 and later

In OpenShift 4.16 and later FIPS clusters, data science pipelines were accessible through the OpenShift AI Dashboard. However, connections to the pipelines API server from the KFP SDK failed due to a TLS handshake error. This issue is now resolved.

RHOAIENG-10129 - Notebook and Ray cluster with matching names causes secret resolution failure

Previously, if you created a notebook and a Ray cluster that had matching names in the same namespace, one controller failed to resolve its secret because the secret already had an owner. This issue is now resolved.

RHOAIENG-7887 - Kueue fails to monitor RayCluster or PyTorchJob resources

Previously, when you created a DataScienceCluster CR with all components enabled, the Kueue component was installed before the Ray component and the Training Operator component. As a result, the Kueue component did not monitor RayCluster or PyTorchJob resources. When a user created RayCluster or PyTorchJob resources, Kueue did not control the admission of those resources. This issue is now resolved.

RHOAIENG-583 (previously documented as RHODS-8921 and RHODS-6373) - You cannot create a pipeline server or start a workbench when cumulative character limit is exceeded

When the cumulative character limit of a data science project name and a pipeline server name exceeded 62 characters, you were unable to successfully create a pipeline server. Similarly, when the cumulative character limit of a data science project name and a workbench name exceeded 62 characters, workbenches failed to start. This issue is now resolved.

Incorrect logo on dashboard after upgrading

Previously, after upgrading from OpenShift AI 2.11 to OpenShift AI 2.12, the dashboard could incorrectly display the Open Data Hub logo instead of the Red Hat OpenShift AI logo. This issue is now resolved.

RHOAIENG-11297 - Authentication failure after pipeline run

Previously, during the execution of a pipeline run, a connection error could occur due to a certificate authentication failure. This certificate authentication failure could be caused by the use of a multi-line string separator for customCABundle in the default-dsci object, which was not supported by data science pipelines. This issue is now resolved.

RHOAIENG-11232 - Distributed workloads: Kueue alerts do not provide runbook link

After a Kueue alert fires, the cluster administrator can click Observe Alerting Alerts and click the name of the alert to open its Alert details page. On the Alert details page, the Runbook section now provides a link to the appropriate runbook to help to diagnose and resolve the issues that triggered the alert. Previously, the runbook link was missing.

RHOAIENG-10665 - Unable to query Speculating with a draft model for granite model

Previously, you could not use speculative decoding on the granite-7b model and granite-7b-accelerator draft model. When querying these models, the queries failed with an internal error. This issue is now resolved.

RHOAIENG-9481 - Pipeline runs menu glitches when clicking action menu

Previously, when you clicked the action menu (⋮) next to a pipeline run on the Experiments > Experiments and runs page, the menu that appeared was not fully visible, and you had to scroll to see all of the menu items. This issue is now resolved.

RHOAIENG-8553 - Workbench created with custom image shows !Deleted flag

Previously, if you disabled the internal image registry on your OpenShift cluster and then created a workbench with a custom image that was imported by using the image tag, for example: quay.io/my-wb-images/my-image:tag, a !Deleted flag was shown in the Notebook image column on the Workbenches tab of the Data Science Projects page. If you stopped the workbench, you could not restart it. This issue is now resolved.

RHOAIENG-6376 - Pipeline run creation fails after setting pip_index_urls in a pipeline component to a URL that contains a port number and path

Previously, when you created a pipeline and set the pip_index_urls value for a component to a URL that contains a port number and path, compiling the pipeline code and then creating a pipeline run could result in an error. This issue is now resolved.

RHOAIENG-4240 - Jobs fail to submit to Ray cluster in unsecured environment

Previously, when running distributed data science workloads from notebooks in an unsecured OpenShift cluster, a ConnectionError: Failed to connect to Ray error message might be shown. This issue is now resolved.

RHOAIENG-9670 - vLLM container intermittently crashes while processing requests

Previously, if you deployed a model by using the vLLM ServingRuntime for KServe runtime on the single-model serving platform and also configured tensor-parallel-size, depending on the hardware platform you used, the kserve-container container would intermittently crash while processing requests. This issue is now resolved.

RHOAIENG-8043 - vLLM errors during generation with mixtral-8x7b

Previously, some models, such as Mixtral-8x7b might have experienced sporadic errors due to a triton issue, such as FileNotFoundError:No such file or directory. This issue is now resolved.

RHOAIENG-2974 - Data science cluster cannot be deleted without its associated initialization object

Previously, you could not delete a DataScienceCluster (DSC) object if its associated DSCInitialization object (DSCI) did not exist. This issue has now been resolved.

RHOAIENG-1205 (previously documented as RHODS-11791) - Usage data collection is enabled after upgrade

Previously, the Allow collection of usage data option would activate whenever you upgraded OpenShift AI. Now, you no longer need to manually deselect the Allow collection of usage data option when you upgrade.

RHOAIENG-1204 (previously documented as ODH-DASHBOARD-1771) - JavaScript error during Pipeline step initializing

Previously, the pipeline Run details page stopped working when a run started. This issue has now been resolved.

RHOAIENG-582 (previously documented as ODH-DASHBOARD-1335) - Rename Edit permission to Contributor

On the Permissions tab for a project, the term Edit has been replaced with Contributor to more accurately describe the actions granted by this permission.

For a complete list of updates, see the Errata advisory.

RHOAIENG-8819 - ibm-granite/granite-3b-code-instruct model fails to deploy on single-model serving platform

Previously, if you tried to deploy the ibm-granite/granite-3b-code-instruct model on the single-model serving platform by using the vLLM ServingRuntime for KServe runtime, the model deployment would fail with an error. This issue is now resolved.

RHOAIENG-8218 - Cannot log in to a workbench created on an OpenShift 4.15 cluster without OCP internal image registry

When you create a workbench on an OpenShift cluster that does not have the OpenShift Container Platform internal image registry enabled, the workbench starts successfully, but you cannot log in to it.

This is a known issue with OpenShift 4.15.x versions earlier than 4.15.15. To resolve this issue, upgrade to OpenShift 4.15.15 or later.

RHOAIENG-7346 - Distributed workloads no longer run from existing pipelines after upgrade

Previously, if you tried to upgrade to OpenShift AI 2.10, a distributed workload would no longer run from an existing pipeline if the cluster was created only inside the pipeline. This issue is now resolved.

RHOAIENG-7209 - Error displays when setting the default pipeline root

Previously, if you tried to set the default pipeline root using the data science pipelines SDK or the OpenShift AI user interface, an error would appear. This issue is now resolved.

RHOAIENG-6711 - ODH-model-controller overwrites the spec.memberSelectors field in ServiceMeshMemberRoll objects

Previously, if you tried to add a project or namespace to a ServiceMeshMemberRoll resource using the spec.memberSelectors field of the ServiceMeshMemberRoll resource, the ODH-model-controller would overwrite the field. This issue is now resolved.

RHOAIENG-6649 - An error is displayed when viewing a model on a model server that has no external route defined

Previously, if you tried to use the dashboard to deploy a model on a model server that did not have external routes enabled, a t.components is undefined error message would appear while the model creation was in progress. This issue is now resolved.

RHOAIENG-3981 - In unsecured environment, the functionality to wait for Ray cluster to be ready gets stuck

Previously, when running distributed data science workloads from notebooks in an unsecured OpenShift cluster, the functionality to wait for the Ray cluster to be ready before proceeding (cluster.wait_ready()) got stuck even when the Ray cluster was ready. This issue is now resolved.

RHOAIENG-2312 - Importing numpy fails in code-server workbench

Previously, if you tried to import numpy, your code-server workbench would fail. This issue is now resolved.

RHOAIENG-1197 - Cannot create pipeline due to the End date picker in the pipeline run creation page defaulting to NaN values when using Firefox on Linux

Previously, if you tried to create a pipeline with a scheduled recurring run using Firefox on Linux, enabling the End Date parameter would result in Not a number (Nan) values for both the date and time. This issue is now resolved.

RHOAIENG-1196 (previously documented as ODH-DASHBOARD-2140) - Package versions displayed in dashboard do not match installed versions

Previously, the dashboard would display inaccurate version numbers for packages such as JupterLab and Notebook. This issue is now resolved.

RHOAIENG-880 - Default pipelines service account is unable to create Ray clusters

Previously, you could not create Ray clusters using the default pipelines Service Account. This issue is now resolved.

RHOAIENG-52 - Token authentication fails in clusters with self-signed certificates

Previously, if you used self-signed certificates, and you used the Python codeflare-sdk in a notebook or in a Python script as part of a pipeline, token authentication would fail. This issue is now resolved.

RHOAIENG-7312 - Model serving fails during query with token authentication in KServe

Previously, if you enabled both the ModelMesh and KServe components in your DataScienceCluster object and added Authorino as an authorization provider, a race condition could occur that resulted in the odh-model-controller pods being rolled out in a state that is appropriate for ModelMesh, but not for KServe and Authorino. In this situation, if you made an inference request to a running model that was deployed using KServe, you saw a 404 - Not Found error. In addition, the logs for the odh-model-controller deployment object showed a Reconciler error message. This issue is now resolved.

RHOAIENG-7181 (previously documented as RHOAIENG-6343)- Some components are set to Removed after installing OpenShift AI

Previously, after you installed OpenShift AI, the managementState field for the codeflare, kueue, and ray components was incorrectly set to Removed instead of Managed in the DataScienceCluster custom resource. This issue is now resolved.

RHOAIENG-7079 (previously documented as RHOAIENG-6317) - Pipeline task status and logs sometimes not shown in OpenShift AI dashboard

Previously, when running pipelines by using Elyra, the OpenShift AI dashboard might not show the pipeline task status and logs, even when the related pods had not been pruned and the information was still available in the OpenShift Console. This issue is now resolved.

RHOAIENG-7070 (previously documented as RHOAIENG-6709) - Jupyter notebook creation might fail when different environment variables specified

Previously, if you started and then stopped a Jupyter notebook, and edited its environment variables in an OpenShift AI workbench, the notebook failed to restart. This issue is now resolved.

RHOAIENG-6853 - Cannot set pod toleration in Elyra pipeline pods

Previously, if you set a pod toleration for an Elyra pipeline pod, the toleration did not take effect. This issue is now resolved.

RHOAIENG-5314 - Data science pipeline server fails to deploy in fresh cluster due to network policies

Previously, if you created a data science pipeline server on a fresh cluster, the user interface remained in a loading state and the pipeline server did not start. This issue is now resolved.

RHOAIENG-4252 - Data science pipeline server deletion process fails to remove ScheduledWorkFlow resource

Previously, the pipeline server deletion process did not remove the ScheduledWorkFlow resource. As a result, new DataSciencePipelinesApplications (DSPAs) did not recognize the redundant ScheduledWorkFlow resource. This issue is now resolved

RHOAIENG-3411 (previously documented as RHOAIENG-3378) - Internal Image Registry is an undeclared hard dependency for Jupyter notebooks spawn process

Previously, before you could start OpenShift AI notebooks and workbenches, you must have already enabled the internal, integrated container image registry in OpenShift. Attempts to start notebooks or workbenches without first enabling the image registry failed with an "InvalidImageName" error. You can now create and use workbenches in OpenShift AI without enabling the internal OpenShift image registry. If you update a cluster to enable or disable the internal image registry, you must recreate existing workbenches for the registry changes to take effect.

RHOAIENG-2541 - KServe controller pod experiences OOM because of too many secrets in the cluster

Previously, if your OpenShift cluster had a large number of secrets, the KServe controller pod could continually crash due to an out-of-memory (OOM) error. This issue is now resolved.

RHOAIENG-1452 - The Red Hat OpenShift AI Add-on gets stuck

Previously, the Red Hat OpenShift AI Add-on uninstall did not delete OpenShift AI components when the install was triggered via OCM APIs. This issue is now resolved.

RHOAIENG-307 - Removing the DataScienceCluster deletes all OpenShift Serverless CRs

Previously, if you deleted the DataScienceCluster custom resource (CR), all OpenShift Serverless CRs (including knative-serving, deployments, gateways, and pods) were also deleted. This issue is now resolved.

RHOAIENG-6709 - Jupyter notebook creation might fail when different environment variables specified

Previously, if you started and then stopped a Jupyter notebook, and edited its environment variables in an OpenShift AI workbench, the notebook failed to restart. This issue is now resolved.

RHOAIENG-6701 - Users without cluster administrator privileges cannot access the job submission endpoint of the Ray dashboard

Previously, users of the distributed workloads feature who did not have cluster administrator privileges for OpenShift might not have been able to access or use the job submission endpoint of the Ray dashboard. This issue is now resolved.

RHOAIENG-6578 - Request without token to a protected inference point not working by default

Previously, if you added Authorino as an authorization provider for the single-model serving platform and enabled token authorization for models that you deployed, it was still possible to query the models without specifying the tokens. This issue is now resolved.

RHOAIENG-6343 - Some components are set to Removed after installing OpenShift AI

Previously, after you installed OpenShift AI, the managementState field for the codeflare, kueue, and ray components was incorrectly set to Removed instead of Managed in the DataScienceCluster custom resource. This issue is now resolved.

RHOAIENG-5067 - Model server metrics page does not load for a model server based on the ModelMesh component

Previously, data science project names that contained capital letters or spaces could cause issues on the model server metrics page for model servers based on the ModelMesh component. The metrics page might not have received data correctly, resulting in a 400 Bad Request error and preventing the page from loading. This issue is now resolved.

RHOAIENG-4966 - Self-signed certificates in a custom CA bundle might be missing from the odh-trusted-ca-bundle configuration map

Previously, if you added a custom certificate authority (CA) bundle to use self-signed certificates, sometimes the custom certificates were missing from the odh-trusted-ca-bundle ConfigMap, or the non-reserved namespaces did not contain the odh-trusted-ca-bundle ConfigMap when the ConfigMap was set to managed. This issue is now resolved.

RHOAIENG-4938 (previously documented as RHOAIENG-4327) - Workbenches do not use the self-signed certificates from centrally configured bundle automatically

There are two bundle options to include self-signed certificates in OpenShift AI, ca-bundle.crt and odh-ca-bundle.crt. Previously, workbenches did not automatically use the self-signed certificates from the centrally configured bundle and you had to define environment variables that pointed to your certificate path. This issue is now resolved.

RHOAIENG-4572- Unable to run data science pipelines after install and upgrade in certain circumstances

Previously, you were unable to run data science pipelines after installing or upgrading OpenShift AI in the following circumstances:

This issue is now resolved.

RHOAIENG-4524 - BuildConfig definitions for RStudio images contain occurrences of incorrect branch

Previously, the BuildConfig definitions for the RStudio and CUDA - RStudio workbench images pointed to the wrong branch in OpenShift AI. This issue is now resolved.

RHOAIENG-3963 - Unnecessary managed resource warning

Previously, when you edited and saved the OdhDashboardConfig custom resource for the redhat-ods-applications project, the system incorrectly displayed a Managed resource warning message. This issue is now resolved.

RHOAIENG-2542 - Inference service pod does not always get an Istio sidecar

Previously, when you deployed a model using the single-model serving platform (which uses KServe), the istio-proxy container could be missing in the resulting pod, even if the inference service had the sidecar.istio.io/inject=true annotation. This issue is now resolved.

RHOAIENG-1666 - The Import Pipeline button is prematurely accessible

Previously, when you imported a pipeline to a workbench that belonged to a data science project, the Import Pipeline button was accessible before the pipeline server was fully available. This issue is now resolved.

RHOAIENG-673 (previously documented as RHODS-12946) - Cannot install from PyPI mirror in disconnected environment or when using private certificates

In disconnected environments, Red Hat OpenShift AI cannot connect to the public-facing PyPI repositories, so you must specify a repository inside your network. Previously, if you were using private TLS certificates and a data science pipeline was configured to install Python packages, the pipeline run would fail. This issue is now resolved.

RHOAIENG-3355 - OVMS on KServe does not use accelerators correctly

Previously, when you deployed a model using the single-model serving platform and selected the OpenVINO Model Server serving runtime, if you requested an accelerator to be attached to your model server, the accelerator hardware was detected but was not used by the model when responding to queries. This issue is now resolved.

RHOAIENG-2869 - Cannot edit existing model framework and model path in a multi-model project

Previously, when you tried to edit a model in a multi-model project using the Deploy model dialog, the Model framework and Path values did not update. This issue is now resolved.

RHOAIENG-2724 - Model deployment fails because fields automatically reset in dialog

Previously, when you deployed a model or edited a deployed model, the Model servers and Model framework fields in the "Deploy model" dialog might have reset to the default state. The Deploy button might have remained enabled even though these mandatory fields no longer contained valid values. This issue is now resolved.

RHOAIENG-2099 - Data science pipeline server fails to deploy in fresh cluster

Previously, when you created a data science pipeline server on a fresh cluster, the user interface remained in a loading state and the pipeline server did not start. This issue is now resolved.

RHOAIENG-1199 (previously documented as ODH-DASHBOARD-1928) - Custom serving runtime creation error message is unhelpful

Previously, when you tried to create or edit a custom model-serving runtime and an error occurred, the error message did not indicate the cause of the error. The error messages have been improved.

RHOAIENG-556 - ServingRuntime for KServe model is created regardless of error

Previously, when you tried to deploy a KServe model and an error occurred, the InferenceService custom resource (CR) was still created and the model was shown in the Data Science Projects page, but the status would always remain unknown. The KServe deploy process has been updated so that the ServingRuntime is not created if an error occurs.

RHOAIENG-548 (previously documented as ODH-DASHBOARD-1776) - Error messages when user does not have project administrator permission

Previously, if you did not have administrator permission for a project, you could not access some features, and the error messages did not explain why. For example, when you created a model server in an environment where you only had access to a single namespace, an Error creating model server error message appeared. However, the model server is still successfully created. This issue is now resolved.

RHOAIENG-66 - Ray dashboard route deployed by CodeFlare SDK exposes self-signed certs instead of cluster cert

Previously, when you deployed a Ray cluster by using the CodeFlare SDK with the openshift_oauth=True option, the resulting route for the Ray cluster was secured by using the passthrough method and as a result, the self-signed certificate used by the OAuth proxy was exposed. This issue is now resolved.

RHOAIENG-12 - Cannot access Ray dashboard from some browsers

In some browsers, users of the distributed workloads feature might not have been able to access the Ray dashboard because the browser automatically changed the prefix of the dashboard URL from http to https. This issue is now resolved.

RHODS-6216 - The ModelMesh oauth-proxy container is intermittently unstable

Previously, ModelMesh pods did not deploy correctly due to a failure of the ModelMesh oauth-proxy container. This issue occurred intermittently and only if authentication was enabled in the ModelMesh runtime environment. This issue is now resolved.

RHOAIENG-535 - Metrics graph showing HTTP requests for deployed models is incorrect if there are no HTTP requests

Previously, if a deployed model did not receive at least one HTTP request for each of the two data types (success and failed), the graphs that show HTTP request performance metrics (for all models on the model server or for the specific model) rendered incorrectly, with a straight line that indicated a steadily increasing number of failed requests. This issue is now resolved.

RHOAIENG-1467 - Serverless net-istio controller pod might hit OOM

Previously, the Knative net-istio-controller pod (which is a dependency for KServe) might continuously crash due to an out-of-memory (OOM) error. This issue is now resolved.

RHOAIENG-1899 (previously documented as RHODS-6539) - The Anaconda Professional Edition cannot be validated and enabled

Previously, you could not enable the Anaconda Professional Edition because the dashboard’s key validation for it was inoperable. This issue is now resolved.

RHOAIENG-2269 - (Single-model) Dashboard fails to display the correct number of model replicas

Previously, on a single-model serving platform, the Models and model servers section of a data science project did not show the correct number of model replicas. This issue is now resolved.

RHOAIENG-2270 - (Single-model) Users cannot update model deployment settings

Previously, you couldn’t edit the deployment settings (for example, the number of replicas) of a model you deployed with a single-model serving platform. This issue is now resolved.

RHODS-8865 - A pipeline server fails to start unless you specify an Amazon Web Services (AWS) Simple Storage Service (S3) bucket resource

Previously, when you created a data connection for a data science project, the AWS_S3_BUCKET field was not designated as a mandatory field. However, if you attempted to configure a pipeline server with a data connection where the AWS_S3_BUCKET field was not populated, the pipeline server failed to start successfully. This issue is now resolved. The Configure pipeline server dialog has been updated to include the Bucket field as a mandatory field.

RHODS-12899 - OpenVINO runtime missing annotation for NVIDIA GPUs

Previously, if a user selected the OpenVINO model server (supports GPUs) runtime and selected an NVIDIA GPU accelerator in the model server user interface, the system could display a unnecessary warning that the selected accelerator was not compatible with the selected runtime. The warning is no longer displayed.

RHOAIENG-84 - Cannot use self-signed certificates with KServe

Previously, the single-model serving platform did not support self-signed certificates. This issue is now resolved. To use self-signed certificates with KServe, follow the steps described in Working with certificates.

RHOAIENG-164 - Number of model server replicas for Kserve is not applied correctly from the dashboard

Previously, when you set a number of model server replicas different from the default (1), the model (server) was still deployed with 1 replica. This issue is now resolved.

RHOAIENG-288 - Recommended image version label for workbench is shown for two versions

Most of the workbench images that are available in OpenShift AI are provided in multiple versions. The only recommended version is the latest version. In Red Hat OpenShift AI 2.4 and 2.5, the Recommended tag was erroneously shown for multiple versions of an image. This issue is now resolved.

RHOAIENG-293 - Deprecated ModelMesh monitoring stack not deleted after upgrading from 2.4 to 2.5

In Red Hat OpenShift AI 2.5, the former ModelMesh monitoring stack was no longer deployed because it was replaced by user workload monitoring. However, the former monitoring stack was not deleted during an upgrade to OpenShift AI 2.5. Some components remained and used cluster resources. This issue is now resolved.

RHOAIENG-343 - Manual configuration of OpenShift Service Mesh and OpenShift Serverless does not work for KServe

If you installed OpenShift Serverless and OpenShift Service Mesh and then installed Red Hat OpenShift AI with KServe enabled, KServe was not deployed. This issue is now resolved.

RHOAIENG-517 - User with edit permissions cannot see created models

A user with edit permissions could not see any created models, unless they were the project owner or had admin permissions for the project. This issue is now resolved.

RHOAIENG-804 - Cannot deploy Large Language Models with KServe on FIPS-enabled clusters

Previously, Red Hat OpenShift AI was not yet fully designed for FIPS. You could not deploy Large Language Models (LLMs) with KServe on FIPS-enabled clusters. This issue is now resolved.

RHOAIENG-908 - Cannot use ModelMesh if KServe was previously enabled and then removed

Previously, when both ModelMesh and KServe were enabled in the DataScienceCluster object, and you subsequently removed KServe, you could no longer deploy new models with ModelMesh. You could continue to use models that were previously deployed with ModelMesh. This issue is now resolved.

RHOAIENG-2184 - Cannot create Ray clusters or distributed workloads

Previously, users could not create Ray clusters or distributed workloads in namespaces where they have admin or edit permissions. This issue is now resolved.

ODH-DASHBOARD-1991 - ovms-gpu-ootb is missing recommended accelerator annotation

Previously, when you added a model server to your project, the Serving runtime list did not show the Recommended serving runtime label for the NVIDIA GPU. This issue is now resolved.

RHOAIENG-807 - Accelerator profile toleration removed when restarting a workbench

Previously, if you created a workbench that used an accelerator profile that in turn included a toleration, restarting the workbench removed the toleration information, which meant that the restart could not complete. A freshly created GPU-enabled workbench might start the first time, but never successfully restarted afterwards because the generated pod remained forever pending. This issue is now resolved.

DATA-SCIENCE-PIPELINES-OPERATOR-294 - Scheduled pipeline run that uses data-passing might fail to pass data between steps, or fail the step entirely

A scheduled pipeline run that uses an S3 object store to store the pipeline artifacts might fail with an error such as the following:

Bad value for --endpoint-url "cp": scheme is missing. Must be of the form http://<hostname>/ or https://<hostname>/

This issue occurred because the S3 object store endpoint was not successfully passed to the pods for the scheduled pipeline run. This issue is now resolved.

RHODS-4769 - GPUs on nodes with unsupported taints cannot be allocated to notebook servers

GPUs on nodes marked with any taint other than the supported nvidia.com/gpu taint could not be selected when creating a notebook server. This issue is now resolved.

RHODS-6346 - Unclear error message displays when using invalid characters to create a data science project

When creating a data science project’s data connection, workbench, or storage connection using invalid special characters, the following error message was displayed:

the object provided is unrecognized (must be of type Secret): couldn't get version/kind; json parse error: unexpected end of JSON input ({"apiVersion":"v1","kind":"Sec ...)

The error message failed to clearly indicate the problem. The error message now indicates that invalid characters were entered.

RHODS-6950 - Unable to scale down workbench GPUs when all GPUs in the cluster are being used

In earlier releases, it was not possible to scale down workbench GPUs if all GPUs in the cluster were being used. This issue applied to GPUs being used by one workbench, and GPUs being used by multiple workbenches. You can now scale down the GPUs by selecting None from the Accelerators list.

RHODS-8939 - Default shared memory for a Jupyter notebook created in a previous release causes a runtime error

Starting with release 1.31, this issue is resolved, and the shared memory for any new notebook is set to the size of the node.

For a Jupyter notebook created in a release earlier than 1.31, the default shared memory for a Jupyter notebook is set to 64 MB and you cannot change this default value in the notebook configuration.

To fix this issue, you must recreate the notebook or follow the process described in the Knowledgebase article How to change the shared memory for a Jupyter notebook in Red Hat OpenShift AI.

RHODS-9030 - Uninstall process for OpenShift AI might become stuck when removing kfdefs resources

The steps for uninstalling the OpenShift AI managed service are described in Uninstalling OpenShift AI.

However, even when you followed this guide, you might have seen that the uninstall process did not finish successfully. Instead, the process stayed on the step of deleting kfdefs resources that were used by the Kubeflow Operator. As shown in the following example, kfdefs resources might exist in the redhat-ods-applications, redhat-ods-monitoring, and rhods-notebooks namespaces:

$ oc get kfdefs.kfdef.apps.kubeflow.org -A

NAMESPACE                  NAME                                   AGE
redhat-ods-applications    rhods-anaconda                         3h6m
redhat-ods-applications    rhods-dashboard                        3h6m
redhat-ods-applications    rhods-data-science-pipelines-operator  3h6m
redhat-ods-applications    rhods-model-mesh                       3h6m
redhat-ods-applications    rhods-nbc                              3h6m
redhat-ods-applications    rhods-osd-config                       3h6m
redhat-ods-monitoring      modelmesh-monitoring                   3h6m
redhat-ods-monitoring      monitoring                             3h6m
rhods-notebooks            rhods-notebooks                        3h6m
rhods-notebooks            rhods-osd-config                       3h5m

Failed removal of the kfdefs resources might have also prevented later installation of a newer version of OpenShift AI. This issue no longer occurs.

RHODS-9764 - Data connection details get reset when editing a workbench

When you edited a workbench that had an existing data connection and then selected the Create new data connection option, the edit page might revert to the Use existing data connection option before you had finished specifying the new connection details. This issue is now resolved.

RHODS-9583 - Data Science dashboard did not detect an existing OpenShift Pipelines installation

When the OpenShift Pipelines Operator was installed as a global operator on your cluster, the OpenShift AI dashboard did not detect it. The OpenShift Pipelines Operator is now detected successfully.

ODH-DASHBOARD-1639 - Wrong TLS value in dashboard route

Previously, when a route was created for the OpenShift AI dashboard on OpenShift, the tls.termination field had an invalid default value of Reencrypt. This issue is now resolved. The new value is reencrypt.

ODH-DASHBOARD-1638 - Name placeholder in Triggered Runs tab shows Scheduled run name

Previously, when you clicked Pipelines > Runs and then selected the Triggered tab to configure a triggered run, the example value shown in the Name field was Scheduled run name. This issue is now resolved.

ODH-DASHBOARD-1547 - "We can’t find that page" message displayed in dashboard when pipeline operator installed in background

Previously, when you used the Data Science Pipelines page of the dashboard to install the OpenShift Pipelines Operator, when the Operator installation was complete, the page refreshed to show a We can't find that page message. This issue is now resolved. When the Operator installation is complete, the dashboard redirects you to the Pipelines page, where you can create a pipeline server.

ODH-DASHBOARD-1545 - Dashboard keeps scrolling to bottom of project when Models tab is expanded

Previously, on the Data Science Projects page of the dashboard, if you clicked the Deployed models tab to expand it and then tried to perform other actions on the page, the page automatically scrolled back to the Deployed models section. This affected your ability to perform other actions. This issue is now resolved.

NOTEBOOKS-156 - Elyra included an example runtime called Test

Previously, Elyra included an example runtime configuration called Test. If you selected this configuration when running a data science pipeline, you could see errors. The Test configuration has now been removed.

RHODS-9622 - Duplicating a scheduled pipeline run does not copy the existing period and pipeline input parameter values

Previously, when you duplicated a scheduled pipeline run that had a periodic trigger, the duplication process did not copy the configured execution frequency for the recurring run or the specified pipeline input parameters. This issue is now resolved.

RHODS-8932 - Incorrect cron format was displayed by default when scheduling a recurring pipeline run

When you scheduled a recurring pipeline run by configuring a cron job, the OpenShift AI interface displayed an incorrect format by default. It now displays the correct format.

RHODS-9374 - Pipelines with non-unique names did not appear in the data science project user interface

If you launched a notebook from a Jupyter application that supported Elyra, or if you used a workbench, when you submitted a pipeline to be run, pipelines with non-unique names did not appear in the Pipelines section of the relevant data science project page or the Pipelines heading of the data science pipelines page. This issue has now been resolved.

RHODS-9329 - Deploying a custom model-serving runtime could result in an error message

Previously, if you used the OpenShift AI dashboard to deploy a custom model-serving runtime, the deployment process could fail with an Error retrieving Serving Runtime message. This issue is now resolved.

RHODS-9064 - After upgrade, the Data Science Pipelines tab was not enabled on the OpenShift AI dashboard

When you upgraded from OpenShift AI 1.26 to OpenShift AI 1.28, the Data Science Pipelines tab was not enabled in the OpenShift AI dashboard. This issue is resolved in OpenShift AI 1.29.

RHODS-9443 - Exporting an Elyra pipeline exposed S3 storage credentials in plain text

In OpenShift AI 1.28.0, when you exported an Elyra pipeline from JupyterLab in Python DSL format or YAML format, the generated output contained S3 storage credentials in plain text. This issue has been resolved in OpenShift AI 1.28.1. However, after you upgrade to OpenShift AI 1.28.1, if your deployment contains a data science project with a pipeline server and a data connection, you must perform the following additional actions for the fix to take effect:

Furthermore, to confirm that your Elyra runtime configuration contains the fix, perform the following actions:

RHODS-8460 - When editing the details of a shared project, the user interface remained in a loading state without reporting an error

When a user with permission to edit a project attempted to edit its details, the user interface remained in a loading state and did not display an appropriate error message. Users with permission to edit projects cannot edit any fields in the project, such as its description. Those users can edit only components belonging to a project, such as its workbenches, data connections, and storage.

The user interface now displays an appropriate error message and does not try to update the project description.

RHODS-8482 - Data science pipeline graphs did not display node edges for running pipelines

If you ran pipelines that did not contain Tekton-formatted Parameters or when expressions in their YAML code, the OpenShift AI user interface did not display connecting edges to and from graph nodes. For example, if you used a pipeline containing the runAfter property or Workspaces, the user interface displayed the graph for the executed pipeline without edge connections. The OpenShift AI user interface now displays connecting edges to and from graph nodes.

RHODS-8923 - Newly created data connections were not detected when you attempted to create a pipeline server

If you created a data connection from within a Data Science project, and then attempted to create a pipeline server, the Configure a pipeline server dialog did not detect the data connection that you created. This issue is now resolved.

RHODS-8461 - When sharing a project with another user, the OpenShift AI user interface text was misleading

When you attempted to share a Data Science project with another user, the user interface text misleadingly implied that users could edit all of its details, such as its description. However, users can edit only components belonging to a project, such as its workbenches, data connections, and storage. This issue is now resolved and the user interface text no longer misleadingly implies that users can edit all of its details.

RHODS-8462 - Users with "Edit" permission could not create a Model Server

Users with "Edit" permissions can now create a Model Server without token authorization. Users must have "Admin" permissions to create a Model Server with token authorization.

RHODS-8796 - OpenVINO Model Server runtime did not have the required flag to force GPU usage

OpenShift AI includes the OpenVINO Model Server (OVMS) model-serving runtime by default. When you configured a new model server and chose this runtime, the Configure model server dialog enabled you to specify a number of GPUs to use with the model server. However, when you finished configuring the model server and deployed models from it, the model server did not actually use any GPUs. This issue is now resolved and the model server uses the GPUs.

RHODS-8861 - Changing the host project when creating a pipeline ran resulted in an inaccurate list of available pipelines

If you changed the host project while creating a pipeline run, the interface failed to make the pipelines of the new host project available. Instead, the interface showed pipelines that belong to the project you initially selected on the Data Science Pipelines > Runs page. This issue is now resolved. You no longer select a pipeline from the Create run page. The pipeline selection is automatically updated when you click the Create run button, based on the current project and its pipeline.

RHODS-8249 - Environment variables uploaded as ConfigMap were stored in Secret instead

Previously, in the OpenShift AI interface, when you added environment variables to a workbench by uploading a ConfigMap configuration, the variables were stored in a Secret object instead. This issue is now resolved.

RHODS-7975 - Workbenches could have multiple data connections

Previously, if you changed the data connection for a workbench, the existing data connection was not released. As a result, a workbench could stay connected to multiple data sources. This issue is now resolved.

RHODS-7948 - Uploading a secret file containing environment variables resulted in double-encoded values

Previously, when creating a workbench in a data science project, if you uploaded a YAML-based secret file containing environment variables, the environment variable values were not decoded. Then, in the resulting OpenShift secret created by this process, the encoded values were encoded again. This issue is now resolved.

RHODS-6429 - An error was displayed when creating a workbench with the Intel OpenVINO or Anaconda Professional Edition images

Previously, when you created a workbench with the Intel OpenVINO or Anaconda Professional Edition images, an error appeared during the creation process. However, the workbench was still successfully created. This issue is now resolved.

RHODS-6372 - Idle notebook culler did not take active terminals into account

Previously, if a notebook image had a running terminal, but no active, running kernels, the idle notebook culler detected the notebook as inactive and stopped the terminal. This issue is now resolved.

RHODS-5700 - Data connections could not be created or connected to when creating a workbench

When creating a workbench, users were unable to create a new data connection, or connect to existing data connections.

RHODS-6281 - OpenShift AI administrators could not access Settings page if an admin group was deleted from cluster

Previously, if a Red Hat OpenShift AI administrator group was deleted from the cluster, OpenShift AI administrator users could no longer access the Settings page on the OpenShift AI dashboard. In particular, the following behavior was seen:

This issue is now resolved.

RHODS-1968 - Deleted users stayed logged in until dashboard was refreshed

Previously, when a user’s permissions for the Red Hat OpenShift AI dashboard were revoked, the user would notice the change only after a refresh of the dashboard page.

This issue is now resolved. When a user’s permissions are revoked, the OpenShift AI dashboard locks the user out within 30 seconds, without the need for a refresh.

RHODS-6384 - A workbench data connection was incorrectly updated when creating a duplicated data connection

When creating a data connection that contained the same name as an existing data connection, the data connection creation failed, but the associated workbench still restarted and connected to the wrong data connection. This issue has been resolved. Workbenches now connect to the correct data connection.

RHODS-6370 - Workbenches failed to receive the latest toleration

Previously, to acquire the latest toleration, users had to attempt to edit the relevant workbench, make no changes, and save the workbench again. Users can now apply the latest toleration change by stopping and then restarting their data science project’s workbench.

RHODS-6779 - Models failed to be served after upgrading from OpenShift AI 1.20 to OpenShift AI 1.21

When upgrading from OpenShift AI 1.20 to OpenShift AI 1.21, the modelmesh-serving pod attempted to pull a non-existent image, causing an image pull error. As a result, models were unable to be served using the model serving feature in OpenShift AI. The odh-openvino-servingruntime-container-v1.21.0-15 image now deploys successfully.

RHODS-5945 - Anaconda Professional Edition could not be enabled in OpenShift AI

Anaconda Professional Edition could not be enabled for use in OpenShift AI. Instead, an InvalidImageName error was displayed in the associated pod’s Events page. Anaconda Professional Edition can now be successfully enabled.

RHODS-5822 - Admin users were not warned when usage exceeded 90% and 100% for PVCs created by data science projects.

Warnings indicating when a PVC exceeded 90% and 100% of its capacity failed to display to admin users for PVCs created by data science projects. Admin users can now view warnings about when a PVC exceeds 90% and 100% of its capacity from the dashboard.

RHODS-5889 - Error message was not displayed if a data science notebook was stuck in "pending" status

If a notebook pod could not be created, the OpenShift AI interface did not show an error message. An error message is now displayed if a data science notebook cannot be spawned.

RHODS-5886 - Returning to the Hub Control Panel dashboard from the data science workbench failed

If you attempted to return to the dashboard from your workbench Jupyter notebook by clicking on File Log Out, you were redirected to the dashboard and remained on a "Logging out" page. Likewise, if you attempted to return to the dashboard by clicking on File Hub Control Panel, you were incorrectly redirected to the Start a notebook server page. Returning to the Hub Control Panel dashboard from the data science workbench now works as expected.

RHODS-6101 - Administrators were unable to stop all notebook servers

OpenShift AI administrators could not stop all notebook servers simultaneously. Administrators can now stop all notebook servers using the Stop all servers button and stop a single notebook by selecting Stop server from the action menu beside the relevant user.

RHODS-5891 - Workbench event log was not clearly visible

When creating a workbench, users could not easily locate the event log window in the OpenShift AI interface. The Starting label under the Status column is now underlined when you hover over it, indicating you can click on it to view the notebook status and the event log.

RHODS-6296 - ISV icons did not render when using a browser other than Google Chrome

When using a browser other than Google Chrome, not all ISV icons under Explore and Resources pages were rendered. ISV icons now display properly on all supported browsers.

RHODS-3182 - Incorrect number of available GPUs was displayed in Jupyter

When a user attempts to create a notebook instance in Jupyter, the maximum number of GPUs available for scheduling was not updated as GPUs are assigned. Jupyter now displays the correct number of GPUs available.

RHODS-5890 - When multiple persistent volumes were mounted to the same directory, workbenches failed to start

When mounting more than one persistent volume (PV) to the same mount folder in the same workbench, creation of the notebook pod failed and no errors were displayed to indicate there was an issue.

RHODS-5768 - Data science projects were not visible to users in Red Hat OpenShift AI

Removing the [DSP] suffix at the end of a project’s Display Name property caused the associated data science project to no longer be visible. It is no longer possible for users to remove this suffix.

RHODS-5701 - Data connection configuration details were overwritten

When a data connection was added to a workbench, the configuration details for that data connection were saved in environment variables. When a second data connection was added, the configuration details are saved using the same environment variables, which meant the configuration for the first data connection was overwritten. At the moment, users can add a maximum of one data connection to each workbench.

RHODS-5252 - The notebook Administration page did not provide administrator access to a user’s notebook server

The notebook Administration page, accessed from the OpenShift AI dashboard, did not provide the means for an administrator to access a user’s notebook server. Administrators were restricted to only starting or stopping a user’s notebook server.

RHODS-2438 - PyTorch and TensorFlow images were unavailable when upgrading

When upgrading from OpenShift AI 1.3 to a later version, PyTorch and TensorFlow images were unavailable to users for approximately 30 minutes. As a result, users were unable to start PyTorch and TensorFlow notebooks in Jupyter during the upgrade process. This issue has now been resolved.

RHODS-5354 - Environment variable names were not validated when starting a notebook server

Environment variable names were not validated on the Start a notebook server page. If an invalid environment variable was added, users were unable to successfully start a notebook. The environmental variable name is now checked in real-time. If an invalid environment variable name is entered, an error message displays indicating valid environment variable names must consist of alphabetic characters, digits, _, -, or ., and must not start with a digit.

RHODS-4617 - The Number of GPUs drop-down was only visible if there were GPUs available

Previously, the Number of GPUs drop-down was only visible on the Start a notebook server page if GPU nodes were available. The Number of GPUs drop-down now also correctly displays if an autoscaling machine pool is defined in the cluster, even if no GPU nodes are currently available, possibly resulting in the provisioning of a new GPU node on the cluster.

RHODS-5420 - Cluster admin did not get administrator access if it was the only user present in the cluster

Previously, when the cluster admin was the only user present in the cluster, it did not get Red Hat OpenShift administrator access automatically. Administrator access is now correctly applied to the cluster admin user.

RHODS-4321 - Incorrect package version displayed during notebook selection

The Start a notebook server page displayed an incorrect version number (11.4 instead of 11.7) for the CUDA notebook image. The version of CUDA installed is no longer specified on this page.

RHODS-5001 - Admin users could add invalid tolerations to notebook pods

An admin user could add invalid tolerations on the Cluster settings page without triggering an error. If a invalid toleration was added, users were unable to successfully start notebooks. The toleration key is now checked in real-time. If an invalid toleration name is entered, an error message displays indicating valid toleration names consist of alphanumeric characters, -, _, or ., and must start and end with an alphanumeric character.

RHODS-5100 - Group role bindings were not applied to cluster administrators

Previously, if you had assigned cluster admin privileges to a group rather than a specific user, the dashboard failed to recognize administrative privileges for users in the administrative group. Group role bindings are now correctly applied to cluster administrators as expected.

RHODS-4947 - Old Minimal Python notebook image persisted after upgrade

After upgrading from OpenShift AI 1.14 to 1.15, the older version of the Minimal Python notebook persisted, including all associated package versions. The older version of the Minimal Python notebook no longer persists after upgrade.

RHODS-4935 - Excessive "missing x-forwarded-access-token header" error messages displayed in dashboard log

The rhods-dashboard pod’s log contained an excessive number of "missing x-forwarded-access-token header" error messages due to a readiness probe hitting the /status endpoint. This issue has now been resolved.

RHODS-2653 - Error occurred while fetching the generated images in the sample Pachyderm notebook

An error occurred when a user attempted to fetch an image using the sample Pachyderm notebook in Jupyter. The error stated that the image could not be found. Pachyderm has corrected this issue.

RHODS-4584 - Jupyter failed to start a notebook server using the OpenVINO notebook image

Jupyter’s Start a notebook server page failed to start a notebook server using the OpenVINO notebook image. Intel has provided an update to the OpenVINO operator to correct this issue.

RHODS-4923 - A non-standard check box displayed after disabling usage data collection

After disabling usage data collection on the Cluster settings page, when a user accessed another area of the OpenShift AI dashboard, and then returned to the Cluster settings page, the Allow collection of usage data check box had a non-standard style applied, and therefore did not look the same as other check boxes when selected or cleared.

RHODS-4938 - Incorrect headings were displayed in the Notebook Images page

The Notebook Images page, accessed from the Settings page on the OpenShift AI dashboard, displayed incorrect headings in the user interface. The Notebook image settings heading displayed as BYON image settings, and the Import Notebook images heading displayed as Import BYON images. The correct headings are now displayed as expected.

RHODS-4818 - Jupyter was unable to display images when the NVIDIA GPU add-on was installed

The Start a notebook server page did not display notebook images after installing the NVIDIA GPU add-on. Images are now correctly displayed, and can be started from the Start a notebook server page.

RHODS-4797 - PVC usage limit alerts were not sent when usage exceeded 90% and 100%

Alerts indicating when a PVC exceeded 90% and 100% of its capacity failed to be triggered and sent. These alerts are now triggered and sent as expected.

RHODS-4366 - Cluster settings were reset on operator restart

When the OpenShift AI operator pod was restarted, cluster settings were sometimes reset to their default values, removing any custom configuration. The OpenShift AI operator was restarted when a new version of OpenShift AI was released, and when the node that ran the operator failed. This issue occurred because the operator deployed ConfigMaps incorrectly. Operator deployment instructions have been updated so that this no longer occurs.

RHODS-4318 - The OpenVINO notebook image failed to build successfully

The OpenVINO notebook image failed to build successfully and displayed an error message. This issue has now been resolved.

RHODS-3743 - Starburst Galaxy quick start did not provide download link in the instruction steps

The Starburst Galaxy quick start, located on the Resources page on the dashboard, required the user to open the explore-data.ipynb notebook, but failed to provide a link within the instruction steps. Instead, the link was provided in the quick start’s introduction.

RHODS-1974 - Changing alert notification emails required pod restart

Changes to the list of notification email addresses in the Red Hat OpenShift AI Add-On were not applied until after the rhods-operator pod and the prometheus-* pod were restarted.

RHODS-2738 - Red Hat OpenShift API Management 1.15.2 add-on installation did not successfully complete

For OpenShift AI installations that are integrated with the Red Hat OpenShift API Management 1.15.2 add-on, the Red Hat OpenShift API Management installation process did not successfully obtain the SMTP credentials secret. Subsequently, the installation did not complete.

RHODS-3237 - GPU tutorial did not appear on dashboard

The "GPU computing" tutorial, located at Gtc2018-numba, did not appear on the Resources page on the dashboard.

RHODS-3069 - GPU selection persisted when GPU nodes were unavailable

When a user provisioned a notebook server with GPU support, and the utilized GPU nodes were subsequently removed from the cluster, the user could not create a notebook server. This occurred because the most recently used setting for the number of attached GPUs was used by default.

RHODS-3181 - Pachyderm now compatible with OpenShift Dedicated 4.10 clusters

Pachyderm was not initially compatible with OpenShift Dedicated 4.10, and so was not available in OpenShift AI running on an OpenShift Dedicated 4.10 cluster. Pachyderm is now available on and compatible with OpenShift Dedicated 4.10.

RHODS-2160 - Uninstall process failed to complete when both OpenShift AI and OpenShift API Management were installed

When OpenShift AI and OpenShift API Management are installed together on the same cluster, they use the same Virtual Private Cluster (VPC). The uninstall process for these Add-ons attempts to delete the VPC. Previously, when both Add-ons are installed, the uninstall process for one service was blocked because the other service still had resources in the VPC. The cleanup process has been updated so that this conflict does not occur.

RHODS-2747 - Images were incorrectly updated after upgrading OpenShift AI

After the process to upgrade OpenShift AI completed, Jupyter failed to update its notebook images. This was due to an issue with the image caching mechanism. Images are now correctly updating after an upgrade.

RHODS-2425 - Incorrect TensorFlow and TensorBoard versions displayed during notebook selection

The Start a notebook server page displayed incorrect version numbers (2.4.0) for TensorFlow and TensorBoard in the TensorFlow notebook image. These versions have been corrected to TensorFlow 2.7.0 and TensorBoard 2.6.0.

RHODS-24339 - Quick start links did not display for enabled applications

For some applications, the Open quick start link failed to display on the application tile on the Enabled page. As a result, users did not have direct access to the quick start tour for the relevant application.

RHODS-2215 - Incorrect Python versions displayed during notebook selection

The Start a notebook server page displayed incorrect versions of Python for the TensorFlow and PyTorch notebook images. Additionally, the third integer of package version numbers is now no longer displayed.

RHODS-1977 - Ten minute wait after notebook server start fails

If the Jupyter leader pod failed while the notebook server was being started, the user could not access their notebook server until the pod restarted, which took approximately ten minutes. This process has been improved so that the user is redirected to their server when a new leader pod is elected. If this process times out, users see a 504 Gateway Timeout error, and can refresh to access their server.