1. Documentation
2. 2025.1 (Latest) 2025.0 2024.1 2024.0 10.3 10.2 10.1 10.0 9.10 9.9 9.8 9.7 9.6 9.5 9.4 9.3 9.2 9.1 9.0 8.2 8.1 8.0 7.6
3. Hub
4. Releases
5. 2025.1.1

What's New in Altair AI Hub 2025.1.1?

Released: 13th June 2025

The following describes the enhancements and bug fixes in AI Hub 2025.1.1:

AI Hub Server

Enhancements

• Web API Agents/Scoring Agents respond with custom set Content-Type annotation for file output as HTTP header when set
• Upgrade to Spring Boot 3.4.6 and move to latest library versions

Fixed

• Fixed an issue that file inputs don't have the correct Content-Type annotation set for endpoint invocation in Web API Agents/Scoring Agents
• Removed the file type check for the upload of queued extensions
• Test button is not clickable now in Endpoints' grouped-by-deployment view, when no agent is online
• DISABLED webapi-deployment state is now translated

License Proxy

• Upgrade to Spring Boot 3.4.6 and move to latest library versions

Deployment fixes and enhancements

Upgrades

• Upgraded Keycloak to 26.2.5
• Upgraded Alpine to 3.22
• Upgraded Python to 3.13.5
• Upgraded Grafana to 11.6.2
• Upgraded Jupyterhub to 5.3.0-7
• Upgraded Activemq to 2.41.0
• Upgraded NVidia Cuda version in our GPU images to 12.9

Enhancements

• Refactored Jupyter Notebook profile definitions in the values.yaml file
• Added an init container to fix ownership and permissions of ActiveMQ data for docker compose deployments and non-OpenShift helm deployments

Fixed

• Refactored docker images to meet OpenShift security requirements so all SecurityContextConstraint related roleBindings and serviceAccounts are removed from the helm chart

Manual upgrade steps required only on OpenShift deployments

Upgrading to 2025.1.1 from previous versions requires changing the ownership and permission of ActiveMQ data .

For non-OpenShift, this is automated using an init container, but not on Openshift this has to be done manually because this requres privileged permissions.

Avoid using the --force flag when running helm upgrade command, there may be leftover objects that can not be deleted during the upgrade, please see next section.

ActiveMQ Artemis data ownership and permission fix

Required only for OpenShift deployment upgrades

Create a file with the name amq-volume-fix.yml

amq-volume-fix.yml: ```yaml apiVersion: v1 kind: ServiceAccount metadata: name: amq-vol-fix-sa namespace:

apiVersion: rbac.authorization.k8s.io/v1 kind: RoleBinding metadata: name: amq-anyuid-scc-binding namespace: subjects: - kind: ServiceAccount name: amq-vol-fix-sa namespace: roleRef: apiGroup: rbac.authorization.k8s.io kind: ClusterRole name: system:openshift:scc:anyuid

apiVersion: batch/v1 kind: Job metadata: name: amq-vol-fix-job namespace: <> spec: template: spec: serviceAccountName: amq-vol-fix-sa securityContext: fsGroup: 0 restartPolicy: OnFailure volumes: - name: activemq-artemis-pv persistentVolumeClaim: claimName: activemq-artemis-pvc containers: - name: amq-volume-fix image: busybox:latest command: ["/bin/sh", "-c"] args: ["chown -R 1001:0 /var/lib/artemis-instance/data && chmod -R 2775 /var/lib/artemis-instance/data"] volumeMounts: - name: activemq-artemis-pv mountPath: /var/lib/artemis-instance/data subPath: activemq-artemis backoffLimit: 3 ```

Than scale down the ActiveMQ deployment, create the objects above with the apply command and then scale the ActiveMQ deployment back up.

# activemq pvc is RWO so activemq deployment needs to be scaled in
oc scale deployment/activemq-dep --replicas=0 -n <namespace>

# Deploy amq-volume-fix.yml file
oc apply -f amq-volume-fix.yml

# Once the job finishes, scale out the activemq deployment
oc scale deployment/activemq-dep --replicas=1 -n <namespace>

If you do the steps above before upgrading to 2025.1.1, your ActiveMQ deployment will not start because the ownership and permissions change until you don't upgrade to 2025.1.1.

If you do the steps above after upgrading to 2025.1.1, your ActiveMQ deployment will not start because the ownership and permissions are not set until you do the steps above.

Remove leftover SecurityContextConstraint related ServiceAccounts and roleBindings before upgrade

Required only for OpenShift deployment upgrades

1. Scale down the relevant components

oc scale deployment/activemq-dep --replicas=0 -n <namespace>
oc scale deployment/aihub-backend-dep --replicas=0 -n <namespace>
oc scale deployment/grafana-proxy-dep --replicas=0 -n <namespace>

1. Manually edit and remove ServiceAccount references from the following components

Remove the spec.template.spec.serviceAccountName field, it is no longer needed

kubectl edit deployment activemq-dep -n <namespace>
kubectl edit deployment aihub-backend-dep -n <namespace>
kubectl edit deployment grafana-proxy-dep -n <namespace>

1. Remove the ServiceAccounts and roleBindings

kubectl delete ServiceAccount activemq-serviceaccount 
kubectl delete ServiceAccount aihub-backend-serviceaccount
kubectl delete ServiceAccount grafana-proxy-serviceaccount

kubectl delete RoleBinding activemq-scc-anyuid-binding
kubectl delete RoleBinding aihub-backend-scc-nonroot-binding
kubectl delete RoleBinding grafana-proxy-scc-node-anyuid-binding

1. You can continue wiith the helm upgrade command after the above steps.