Critical NVIDIA DPF Documentation Fixes for v26.4.1 - #117440
Conversation
openshift-ci
Bot
added
the
size/XXL
| :_mod-docs-content-type: PROCEDURE | ||
|
|
||
| [id="nw-dpf-installing-grafana-for-dts_{context}"] | ||
| = Installing Grafana for DTS metrics visualization |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.ShortDescription: Assign [role="_abstract"] to a paragraph to use it as in DITA.
| + | ||
| .. In Grafana, go to **Dashboards** → **"DOCA DPU Telemetry (DTS)"** | ||
| .. Use the **DPU (source)** dropdown at the top to focus on specific DPUs or select **All** | ||
| .. Adjust the time range (top-right corner) - the dashboard refreshes every 30 seconds automatically |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'upper left', 'upper right', 'upper-left', or 'upper-right' rather than 'top-right'. For more information, see RedHat.TermsErrors.
| :_mod-docs-content-type: REFERENCE | ||
|
|
||
| [id="nw-dpf-release-notes_{context}"] | ||
| = DPF Operator release notes and troubleshooting |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.ShortDescription: Assign [role="_abstract"] to a paragraph to use it as in DITA.
|
|
||
| == Troubleshooting | ||
|
|
||
| === DPU provisioning issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| oc -n dpf-operator-system exec deploy/dpf-operator-controller-manager -- /dpfctl describe dpudeployments | ||
| ---- | ||
|
|
||
| === DPU object state issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| + | ||
| Approve any pending CSRs related to DPU nodes. | ||
|
|
||
| === Management cluster node readiness issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| + | ||
| Ensure all OVN-K pods are running on DPU worker nodes. | ||
|
|
||
| === Grafana deployment issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| effect: NoSchedule | ||
| ---- | ||
|
|
||
| === DTS metrics collection issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| + | ||
| Access the Prometheus web console and navigate to **Status** → **Targets** to verify DTS endpoints are listed and healthy. | ||
|
|
||
| === Certificate and BMC issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
|
|
||
| . Ensure the BareMetalHost resource has correct BMC credentials and connection details. | ||
|
|
||
| === Worker node CSR approval issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
|
|
||
| . Verify the `br-dpu` bridge is correctly configured on worker nodes. | ||
|
|
||
| === DPU node status issues |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
|
|
||
| * link:https://github.com/NVIDIA/doca-platform/blob/public-release-v25.4/docs/public/user-guides/[NVIDIA DPF User Guides] | ||
| * link:https://docs.nvidia.com/networking/display/dpf2504/troubleshooting[NVIDIA DPF Troubleshooting Guide] | ||
| * {product-title} documentation for monitoring and observability |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.RelatedLinks: Content other than links cannot be mapped to DITA related-links.
1239a79 to
739a755
Compare
| :_mod-docs-content-type: PROCEDURE | ||
|
|
||
| [id="nw-dpf-installing-grafana-for-dts_{context}"] | ||
| = Installing Grafana for DTS metrics visualization |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.ShortDescription: Assign [role="_abstract"] to a paragraph to use it as in DITA.
|
|
||
| . Navigate to the DTS dashboard: | ||
| + | ||
| In Grafana, go to *Dashboards* and open *DOCA DPU Telemetry (DTS)*. Use the *DPU (source)* dropdown at the top to focus on a specific DPU or select *All*. Adjust the time range in the top-right corner. The dashboard refreshes every 30 seconds automatically. |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'upper left', 'upper right', 'upper-left', or 'upper-right' rather than 'top-right'. For more information, see RedHat.TermsErrors.
| :_mod-docs-content-type: REFERENCE | ||
|
|
||
| [id="nw-dpf-release-notes_{context}"] | ||
| = DPF Operator release notes |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.ShortDescription: Assign [role="_abstract"] to a paragraph to use it as in DITA.
| * NVIDIA BlueField-3 DPUs | ||
| * Red{nbsp}Hat Enterprise Linux CoreOS (RHCOS) 10.2+ | ||
|
|
||
| === New features and enhancements |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| **Enhanced observability with DTS integration**:: | ||
| Added comprehensive DPU telemetry monitoring through the DOCA Telemetry Service (DTS) with built-in {product-title} Console dashboard integration. DTS metrics are now accessible directly through the {product-title} web console without requiring additional tools. | ||
|
|
||
| **Improved HyperShift integration**:: |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'hosted control planes' rather than 'HyperShift'. For more information, see RedHat.TermsErrors.
| **Enhanced troubleshooting capabilities**:: | ||
| Expanded diagnostic tools and troubleshooting procedures covering DPU provisioning, hosted cluster management, networking issues, and comprehensive log collection. | ||
|
|
||
| === Bug fixes |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| **Networking stability improvements**:: | ||
| Fixed OVN-Kubernetes integration issues that could cause worker nodes to remain in `NotReady` state. | ||
|
|
||
| === Technology Preview features |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
| **DOCA Platform Framework**:: | ||
| The DPF Operator and all related DPU acceleration features are provided as Technology Preview. Technology Preview features are not supported with Red{nbsp}Hat production service level agreements (SLAs) and might not be functionally complete. | ||
|
|
||
| === Deprecated and removed features |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.NestedSection: Level 2, 3, 4, and 5 sections (=== and deeper) are not supported in DITA.
|
|
||
| When opening a support case, collect the following information: | ||
|
|
||
| .Environment information |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| * Hardware specifications (server model, DPU model, firmware versions) | ||
| * Network topology and configuration | ||
|
|
||
| .Configuration files |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| * Network policies and configurations | ||
| * Environment variables used during installation | ||
|
|
||
| .Log files |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| * Kubernetes event logs related to DPF resources | ||
| * Application logs for affected services | ||
|
|
||
| .Status outputs |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| + | ||
| Ensure that the firmware version is compatible with the DOCA Platform Framework. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| + | ||
| Ensure that the firmware version is compatible with the DOCA Platform Framework. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.TaskTitle: Unsupported titles cannot be mapped to DITA tasks.
|
|
||
| If DPU provisioning does not start immediately after adding worker nodes to the cluster, use the following diagnostic steps to identify the root cause. | ||
|
|
||
| .Procedure |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
|
|
||
| If DPU objects remain in the `DPU Cluster Config` state for an extended period, the issue is typically related to pending CSRs in the hosted cluster. | ||
|
|
||
| .Procedure |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
|
|
||
| If management cluster nodes do not become `Ready` after DPU provisioning is complete, the issue is typically related to OVN-Kubernetes CNI configuration. | ||
|
|
||
| .Procedure |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| == Additional resources | ||
|
|
||
| * link:https://docs.nvidia.com/networking/display/dpf2504/[NVIDIA DPF Documentation] | ||
| * {product-title} monitoring and logging documentation |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.RelatedLinks: Content other than links cannot be mapped to DITA related-links.
|
|
||
| * link:https://docs.nvidia.com/networking/display/dpf2504/[NVIDIA DPF Documentation] | ||
| * {product-title} monitoring and logging documentation | ||
| * {product-title} networking troubleshooting procedures |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.RelatedLinks: Content other than links cannot be mapped to DITA related-links.
| $ oc cluster-info | ||
| ---- | ||
| + | ||
| If the cluster is not accessible, check the HyperShift hosted control plane status on the management cluster. |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'hosted control planes' rather than 'HyperShift'. For more information, see RedHat.TermsErrors.
| + | ||
| If the cluster is not accessible, check the HyperShift hosted control plane status on the management cluster. | ||
|
|
||
| . Switch to the management cluster and check the HyperShift components: |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'hosted control planes' rather than 'HyperShift'. For more information, see RedHat.TermsErrors.
| $ oc get pods -n clusters-$HOSTED_CLUSTER_NAME | ||
| ---- | ||
| + | ||
| Verify that all HyperShift control plane pods are running: |
There was a problem hiding this comment.
🤖 [error] RedHat.TermsErrors: Use 'hosted control planes' rather than 'HyperShift'. For more information, see RedHat.TermsErrors.
| + | ||
| Ensure that OVN-Kubernetes pods are running on the DPU ARM cores. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.TaskTitle: Unsupported titles cannot be mapped to DITA tasks.
| + | ||
| Ensure that OVN-Kubernetes pods are running on the DPU ARM cores. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| + | ||
| Verify that SR-IOV policies are correctly applied to DPU-enabled worker nodes. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.BlockTitle: Block titles can only be assigned to examples, figures, and tables in DITA.
| + | ||
| Verify that SR-IOV policies are correctly applied to DPU-enabled worker nodes. | ||
|
|
||
| .Troubleshooting tips |
There was a problem hiding this comment.
🤖 [error] AsciiDocDITA.TaskTitle: Unsupported titles cannot be mapped to DITA tasks.
- Remove DPUServiceTemplate YAML from OVN, HBN, DTS service docs (templates are now auto-managed by dpf-hcp-provisioner-operator) - Fix OVN DPUServiceConfiguration: add hostNetworkNamespace, imagePullSecretName, remove ovnMultiNetworkEnable - Update DPUFlavor OVS scripts with idempotent helpers and conditional restart (both 1500 and 9000 MTU variants) - Update ArgoCD: remove skip-injection labels, Redis patch, add resourceExclusions for PackageManifest - Remove manual ConfigMaps from OVN DPU-Host mode doc (now auto-configured by provisioning controller) - Remove manual worker MachineConfig (now auto-created by dpf-hcp-provisioner-operator), keep MachineConfigPool only - Update DPFHCPProvisioner: replace virtualIP with controlPlaneAvailabilityPolicy - Update DTS ServiceMonitor: use matchExpressions with relabelings instead of matchLabels Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
b746a20 to
a51c86a
Compare
|
@szigmon: The following tests failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here. |
3 participants
Critical NVIDIA DPF Documentation Fixes for v26.4.1
Overview
This PR addresses critical production-blocking issues in NVIDIA DPF operator documentation and adds essential fixes validated against DPF v26.4.1-beta.1.
Key Changes
🚨 Critical Fix: Remove dpf-pull-secret References
Problem: Documentation incorrectly requires
dpf-pull-secretcausing installation failuresRoot Cause: Based on outdated examples that assumed private registries
Evidence:
imagePullSecretsFiles Fixed:
modules/nw-dpf-creating-hbn-service.adoc- RemovedimagePullSecretssectionmodules/nw-dpf-creating-dts-service.adoc- RemovedimagePullSecretssectionmodules/nw-dpf-creating-ovnk-service.adoc- RemovedimagePullSecretNameparameter🔧 Production Grafana Installation Fixes
Problem: "Application is not available" errors in DTS Grafana deployment
Solution: Added complete Grafana installation with control-plane node scheduling
New Module:
modules/nw-dpf-installing-grafana-for-dts.adoc(333 lines)📋 Comprehensive Release Notes
New Module:
modules/nw-dpf-release-notes.adoc(120 lines)🔄 Version Updates
Validation
Real-World Testing
✅ Full DPF v26.4.1-beta.1 deployment tested without dpf-pull-secret
✅ All operator pods running without authentication errors
✅ Grafana fixes validated in production environment
✅ YAML syntax validated with
oc apply --dry-run=clientComparison with Tech Writer PR
This complements PR #117142 by adding critical production fixes not covered:
Impact
Files Changed
Commits
NVIDIA-918: Add NVIDIA DPF Operator documentation- Base documentationAdd critical DTS Grafana installation and production fixes- Production enhancementsRemove dpf-pull-secret references from DPU service configurations- Critical fixThis PR delivers production-ready documentation that actually works for enterprise deployments.