[Bug]: GPU Operator MPS config-manager cannot signal MPS daemon due to process-target mismatch

Open

#2,970 opened on Sep 25, 2026 1 comment 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty: 3/5
Estimated time: 1-2 days
Newbie friendliness: 68/100

Issue type: Bug
Clarity: Clearly specified
Activity status: Active
Tech stack: go, kubernetes
Domain: devops, infrastructure

Research direction

Start with the MPS DaemonSet template and the config-manager implementation used by the v0.20.1 device-plugin image, comparing the mps-control-daemon command with PROCESS_TO_SIGNAL and its exact cmdline matching. Reproduce the profile switch with the supplied kubectl label command, then verify that config-manager signals the daemon successfully and the MPS DaemonSet returns to 2/2 Ready.

Written by the indexing model from the issue text.

Description

bug needs-triage

Describe the bug
With GPU Operator v26.7.0 and v26.7.1 (other versions not tested), the MPS control-daemon DaemonSet config-manager sidecar cannot reload the MPS control daemon after a device-plugin configuration change.

The MPS DaemonSet template starts the daemon using: command: [mps-control-daemon] but config-manager is configured to search for: PROCESS_TO_SIGNAL: /usr/bin/mps-control-daemon. The config-manager implementation requires an exact match against cmdline[0]. It cannot find the process, exits with code 1, and enters CrashLoopBackOff.

This leaves the MPS DaemonSet at 1/2 ready and the GPU Operator ClusterPolicy in notReady, even when the MPS main container is eventually serving shared-GPU resources.

To Reproduce

Install GPU Operator v26.7.1 with a device-plugin ConfigMap containing two profiles: i.e. default without MPS (full GPU allocation) and one with MPS enabled. The values I used:


mig:
  strategy: none

devicePlugin:
  config:
    create: true
    name: nvidia-device-plugin-config
    default: fullGpu
    data:
      fullGpu: |-
        version: v1
        flags:
          migStrategy: none
      mps-profile: |-
        version: v1
        sharing:
          mps:
            renameByDefault: true
            resources:
              - name: nvidia.com/gpu
                replicas: 4
  mps:
    root: /run/nvidia/mps

Allow daemons to start with default profile:
Change the node to the MPS profile:

kubectl label node <gpu-node> nvidia.com/device-plugin.config=mps-profile --overwrite

Check the MPS config-manager logs:

POD=$(kubectl get pods -n nvidia -l app=nvidia-device-plugin-mps-control-daemon -o jsonpath='{.items[0].metadata.name}')
kubectl logs -n nvidia "$POD" -c config-manager --tail=100

Actual result

Label change detected: nvidia.com/device-plugin.config=mps-profile
Updating to config: mps-profile
Successfully updated to config: mps-profile
Sending signal 'hangup' to '/usr/bin/mps-control-daemon'
error finding pid: no process found

The config-manager sidecar enters CrashLoopBackOff, the MPS DaemonSet remains 1/2, and ClusterPolicy stays notReady.

Expected behavior
The config-manager sends SIGHUP successfully, the MPS control daemon reloads the new profile, and the MPS DaemonSet remains 2/2 Ready.

Environment (please provide the following information):

GPU Operator Version: v26.7.1
OS: Ubuntu24.04
Kernel Version: 6.8.0-generic
Container Runtime Version: containerd 2.2.2-1~~ubuntu.24.04~~noble
Kubernetes Distro and Version: Kubernetes: v1.35.3

Information to attach (optional if deemed irrelevant)

kubernetes pods status: kubectl get pods -n OPERATOR_NAMESPACE

# Other pods are running and completely healthy except the mps-control-daemon
...
nvidia-device-plugin-mps-control-daemon-wqhfq                     1/2     CrashLoopBackOff   12 (106s ago)   39m
...

kubernetes daemonset status: kubectl get ds -n OPERATOR_NAMESPACE

NAME                                      DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR                                                          AGE
nvidia-device-plugin-mps-control-daemon   1         1         0       1            0           nvidia.com/gpu.deploy.device-plugin=true,nvidia.com/mps.capable=true   3h36m

If a pod/ds is in an error state or pending state kubectl describe pod -n OPERATOR_NAMESPACE POD_NAME

 $ kubectl describe pod -n nvidia nvidia-device-plugin-mps-control-daemon-8vqms
 Name:                 nvidia-device-plugin-mps-control-daemon-8vqms
 Namespace:            nvidia
 Priority:             2000001000
 Priority Class Name:  system-node-critical
 Runtime Class Name:   nvidia
 Service Account:      nvidia-device-plugin
 Node:                 k8s-worker-gpu-0/172.22.4.22
 Start Time:           Fri, 25 Sep 2026 19:12:17 +0300
 Labels:               app=nvidia-device-plugin-mps-control-daemon
                       app.kubernetes.io/managed-by=gpu-operator
                       controller-revision-hash=956ddf497
                       helm.sh/chart=gpu-operator-v26.7.1
                       pod-template-generation=8
                       topology.kubernetes.io/region=RegionOne
                       topology.kubernetes.io/zone=nova
 Annotations:          <none>
 Status:               Running
 IP:                   10.244.221.17
 IPs:
   IP:           10.244.221.17
 Controlled By:  DaemonSet/nvidia-device-plugin-mps-control-daemon
 Init Containers:
   toolkit-validation:
     Container ID:  containerd://828c8ff9e644722a867b3df8d520a65dc64dea80a1be2770f9cc8419e4d3fa5c
     Image:         nvcr.io/nvidia/gpu-operator:v26.7.1
     Image ID:      nvcr.io/nvidia/gpu-operator@sha256:b844f9c033e748b8f9925f791f426a531ee8a855f04d4ccea365378634bcefe5
     Port:          <none>
     Host Port:     <none>
     Command:
       sh
       -c
     Args:
       until [ -f /run/nvidia/validations/toolkit-ready ] && { grep -q '^nvidia ' /proc/modules || [ -e /dev/dxg ]; }; do echo waiting for nvidia container stack to be setup; sleep 5; done
     State:          Terminated
       Reason:       Completed
       Exit Code:    0
       Started:      Fri, 25 Sep 2026 19:12:17 +0300
       Finished:     Fri, 25 Sep 2026 19:12:17 +0300
     Ready:          True
     Restart Count:  0
     Environment:    <none>
     Mounts:
       /run/nvidia from run-nvidia (rw)
       /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6xkw (ro)
   mps-control-daemon-mounts:
     Container ID:  containerd://6b72cb1e4350fd323d4a85bb8522412252f321be8868c488b923d5e4663c405b
     Image:         nvcr.io/nvidia/k8s-device-plugin:v0.20.1
     Image ID:      nvcr.io/nvidia/k8s-device-plugin@sha256:27c1b2553a690ca3d29889dcd61df97525d46f58272c0be8dc63c2b18f120075
     Port:          <none>
     Host Port:     <none>
     Command:
       mps-control-daemon
       mount-shm
     State:          Terminated
       Reason:       Completed
       Exit Code:    0
       Started:      Fri, 25 Sep 2026 19:12:18 +0300
       Finished:     Fri, 25 Sep 2026 19:12:18 +0300
     Ready:          True
     Restart Count:  0
     Environment:    <none>
     Mounts:
       /mps from mps-root (rw)
       /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6xkw (ro)
   config-manager-init:
     Container ID:  containerd://e81e9d6ce3c5fca148f959fa6558c661618d4520e301270f373f2314f40d5e82
     Image:         nvcr.io/nvidia/k8s-device-plugin:v0.20.1
     Image ID:      nvcr.io/nvidia/k8s-device-plugin@sha256:27c1b2553a690ca3d29889dcd61df97525d46f58272c0be8dc63c2b18f120075
     Port:          <none>
     Host Port:     <none>
     Command:
       config-manager
     State:          Terminated
       Reason:       Completed
       Exit Code:    0
       Started:      Fri, 25 Sep 2026 19:12:19 +0300
       Finished:     Fri, 25 Sep 2026 19:12:19 +0300
     Ready:          True
     Restart Count:  0
     Environment:
       ONESHOT:              true
       KUBECONFIG:
       NODE_NAME:             (v1:spec.nodeName)
       NODE_LABEL:           nvidia.com/device-plugin.config
       CONFIG_FILE_SRCDIR:   /available-configs
       CONFIG_FILE_DST:      /config/config.yaml
       DEFAULT_CONFIG:       fullGpu
       SEND_SIGNAL:          false
       SIGNAL:
       PROCESS_TO_SIGNAL:
       FALLBACK_STRATEGIES:  empty
     Mounts:
       /available-configs from nvidia-gpu-operator-device-plugin-config (rw)
       /config from config (rw)
       /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6xkw (ro)
 Containers:
   mps-control-daemon-ctr:
     Container ID:  containerd://7fe7a1e57c4cc978869e6da1cff72439c0241c236c3826550fb6b816d0409c20
     Image:         nvcr.io/nvidia/k8s-device-plugin:v0.20.1
     Image ID:      nvcr.io/nvidia/k8s-device-plugin@sha256:27c1b2553a690ca3d29889dcd61df97525d46f58272c0be8dc63c2b18f120075
     Port:          <none>
     Host Port:     <none>
     Command:
       mps-control-daemon
     State:          Running
       Started:      Fri, 25 Sep 2026 19:12:20 +0300
     Ready:          True
     Restart Count:  0
     Environment:
       NODE_NAME:                    (v1:spec.nodeName)
       NVIDIA_VISIBLE_DEVICES:      all
       NVIDIA_DRIVER_CAPABILITIES:  compute,utility
       CONFIG_FILE:                 /config/config.yaml
       MIG_STRATEGY:                none
     Mounts:
       /available-configs from nvidia-gpu-operator-device-plugin-config (rw)
       /config from config (rw)
       /dev/shm from mps-shm (rw)
       /mps from mps-root (rw)
       /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6xkw (ro)
   config-manager:
     Container ID:  containerd://83d50e253dd1b7c1393fe4188b6f20fc9f82786190234ee6128fb9ec25586141
     Image:         nvcr.io/nvidia/k8s-device-plugin:v0.20.1
     Image ID:      nvcr.io/nvidia/k8s-device-plugin@sha256:27c1b2553a690ca3d29889dcd61df97525d46f58272c0be8dc63c2b18f120075
     Port:          <none>
     Host Port:     <none>
     Command:
       config-manager
     State:          Terminated
       Reason:       Error
       Exit Code:    1
       Started:      Fri, 25 Sep 2026 19:12:34 +0300
       Finished:     Fri, 25 Sep 2026 19:12:34 +0300
     Last State:     Terminated
       Reason:       Error
       Exit Code:    1
       Started:      Fri, 25 Sep 2026 19:12:21 +0300
       Finished:     Fri, 25 Sep 2026 19:12:21 +0300
     Ready:          False
     Restart Count:  2
     Environment:
       ONESHOT:              false
       KUBECONFIG:
       NODE_NAME:             (v1:spec.nodeName)
       NODE_LABEL:           nvidia.com/device-plugin.config
       CONFIG_FILE_SRCDIR:   /available-configs
       CONFIG_FILE_DST:      /config/config.yaml
       DEFAULT_CONFIG:       fullGpu
       SEND_SIGNAL:          true
       SIGNAL:               1
       PROCESS_TO_SIGNAL:    /usr/bin/mps-control-daemon
       FALLBACK_STRATEGIES:  empty
     Mounts:
       /available-configs from nvidia-gpu-operator-device-plugin-config (rw)
       /config from config (rw)
       /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6xkw (ro)
 Conditions:
   Type                        Status
   PodReadyToStartContainers   True
   Initialized                 True
   Ready                       False
   ContainersReady             False
   PodScheduled                True
 Volumes:
   run-nvidia:
     Type:          HostPath (bare host directory volume)
     Path:          /run/nvidia
     HostPathType:  Directory
   mps-root:
     Type:          HostPath (bare host directory volume)
     Path:          /run/nvidia/mps
     HostPathType:  DirectoryOrCreate
   mps-shm:
     Type:          HostPath (bare host directory volume)
     Path:          /run/nvidia/mps/shm
     HostPathType:
   nvidia-gpu-operator-device-plugin-config:
     Type:      ConfigMap (a volume populated by a ConfigMap)
     Name:      nvidia-gpu-operator-device-plugin-config
     Optional:  false
   config:
     Type:       EmptyDir (a temporary directory that shares a pod's lifetime)
     Medium:
     SizeLimit:  <unset>
   kube-api-access-z6xkw:
     Type:                    Projected (a volume that contains injected data from multiple sources)
     TokenExpirationSeconds:  3607
     ConfigMapName:           kube-root-ca.crt
     Optional:                false
     DownwardAPI:             true
 QoS Class:                   BestEffort
 Node-Selectors:              nvidia.com/gpu.deploy.device-plugin=true
                              nvidia.com/mps.capable=true
 Tolerations:                 evo.company/nvidia-gpu:NoSchedule op=Exists
                              node.kubernetes.io/disk-pressure:NoSchedule op=Exists
                              node.kubernetes.io/memory-pressure:NoSchedule op=Exists
                              node.kubernetes.io/not-ready:NoExecute op=Exists
                              node.kubernetes.io/pid-pressure:NoSchedule op=Exists
                              node.kubernetes.io/unreachable:NoExecute op=Exists
                              node.kubernetes.io/unschedulable:NoSchedule op=Exists
                              nvidia.com/gpu:NoSchedule op=Exists
 Events:
   Type     Reason     Age                From               Message
   ----     ------     ----               ----               -------
   Normal   Scheduled  43s                default-scheduler  Successfully assigned nvidia/nvidia-device-plugin-mps-control-daemon-8vqms to k8s-worker-gpu-0
   Normal   Pulled     43s                kubelet            spec.initContainers{toolkit-validation}: Container image "nvcr.io/nvidia/gpu-operator:v26.7.1" already present on machine and can be accessed by the pod
   Normal   Created    43s                kubelet            spec.initContainers{toolkit-validation}: Container created
   Normal   Started    43s                kubelet            spec.initContainers{toolkit-validation}: Container started
   Normal   Pulled     42s                kubelet            spec.initContainers{mps-control-daemon-mounts}: Container image "nvcr.io/nvidia/k8s-device-plugin:v0.20.1" already present on machine and can be accessed by the pod
   Normal   Created    42s                kubelet            spec.initContainers{mps-control-daemon-mounts}: Container created
   Normal   Started    42s                kubelet            spec.initContainers{mps-control-daemon-mounts}: Container started
   Normal   Pulled     41s                kubelet            spec.initContainers{config-manager-init}: Container image "nvcr.io/nvidia/k8s-device-plugin:v0.20.1" already present on machine and can be accessed by the pod
   Normal   Created    41s                kubelet            spec.initContainers{config-manager-init}: Container created
   Normal   Started    41s                kubelet            spec.initContainers{config-manager-init}: Container started
   Normal   Pulled     40s                kubelet            spec.containers{mps-control-daemon-ctr}: Container image "nvcr.io/nvidia/k8s-device-plugin:v0.20.1" already present on machine and can be accessed by the pod
   Normal   Created    40s                kubelet            spec.containers{mps-control-daemon-ctr}: Container created
   Normal   Started    40s                kubelet            spec.containers{mps-control-daemon-ctr}: Container started
   Normal   Pulled     26s (x3 over 40s)  kubelet            spec.containers{config-manager}: Container image "nvcr.io/nvidia/k8s-device-plugin:v0.20.1" already present on machine and can be accessed by the pod
   Normal   Created    26s (x3 over 40s)  kubelet            spec.containers{config-manager}: Container created
   Normal   Started    26s (x3 over 40s)  kubelet            spec.containers{config-manager}: Container started
   Warning  BackOff    26s (x2 over 38s)  kubelet            spec.containers{config-manager}: Back-off restarting failed container config-manager in pod nvidia-device-plugin-mps-control-daemon-8vqms_nvidia(2956c3ac-7f02-4752-a362-2950f79074be)

If a pod/ds is in an error state or pending state kubectl logs -n OPERATOR_NAMESPACE POD_NAME --all-containers

 kubectl logs -n nvidia nvidia-device-plugin-mps-control-daemon-8vqms --all-containers
 I0925 16:12:18.620956   72937 main.go:80] "NVIDIA MPS Control Daemon" version=<
 	487d7a93
 	commit: 487d7a93894b0f3ad08097169593972eb3ca8206
  >
 W0925 16:12:19.627450   73059 client_config.go:683] Neither --kubeconfig nor --master was specified.  Using the inClusterConfig.  This might not work.
 I0925 16:12:19.627665   73059 main.go:248] Waiting for change to 'nvidia.com/device-plugin.config' label
 I0925 16:12:19.627672   73059 main.go:250] Label change detected: nvidia.com/device-plugin.config=
 I0925 16:12:19.627697   73059 main.go:362] No value set. Selecting default name: fullGpu
 I0925 16:12:19.627701   73059 main.go:306] Updating to config: fullGpu
 I0925 16:12:19.627733   73059 main.go:321] Successfully updated to config: fullGpu
 I0925 16:12:20.651387   73188 main.go:80] "NVIDIA MPS Control Daemon" version=<
 	487d7a93
 	commit: 487d7a93894b0f3ad08097169593972eb3ca8206
  >
 I0925 16:12:20.651458   73188 main.go:109] Starting OS watcher.
 I0925 16:12:20.651602   73188 main.go:123] Starting Daemons.
 I0925 16:12:20.651616   73188 main.go:166] Loading configuration.
 I0925 16:12:20.651802   73188 main.go:181] Updating config with default resource matching patterns.
 I0925 16:12:20.651860   73188 main.go:192]
 Running with config:
 {
   "version": "v1",
   "flags": {
     "migStrategy": "none",
     "failOnInitError": null,
     "gdrcopyEnabled": null,
     "gdsEnabled": null,
     "mofedEnabled": null,
     "useNodeFeatureAPI": null,
     "deviceDiscoveryStrategy": null,
     "plugin": {
       "passDeviceSpecs": null,
       "deviceListStrategy": null,
       "deviceIDStrategy": null,
       "cdiAnnotationPrefix": null,
       "nvidiaCTKPath": null,
       "containerDriverRoot": null,
       "sharedDevicesAllocationPolicy": null
     }
   },
   "resources": {
     "gpus": [
       {
         "pattern": "*",
         "name": "nvidia.com/gpu"
       }
     ]
   },
   "sharing": {
     "timeSlicing": {}
   },
   "imex": {}
 }
 I0925 16:12:20.651866   73188 main.go:196] Retrieving MPS daemons.
 I0925 16:12:20.651875   73188 manager.go:66] "Sharing strategy is not MPS; skipping MPS manager creation" strategy="none"
 I0925 16:12:20.651887   73188 main.go:205] No devices are configured for MPS sharing; Waiting indefinitely.
 W0925 16:15:10.260092   74999 client_config.go:683] Neither --kubeconfig nor --master was specified.  Using the inClusterConfig.  This might not work.
 I0925 16:15:10.260297   74999 main.go:248] Waiting for change to 'nvidia.com/device-plugin.config' label
 I0925 16:15:10.260304   74999 main.go:250] Label change detected: nvidia.com/device-plugin.config=
 I0925 16:15:10.260326   74999 main.go:362] No value set. Selecting default name: fullGpu
 I0925 16:15:10.260331   74999 main.go:306] Updating to config: fullGpu
 I0925 16:15:10.260374   74999 main.go:321] Successfully updated to config: fullGpu
 I0925 16:15:10.260378   74999 main.go:325] Sending signal 'hangup' to '/usr/bin/mps-control-daemon'
 E0925 16:15:10.264400   74999 main.go:210] error finding pid: no process found

Output from running nvidia-smi from the driver container: kubectl exec DRIVER_POD_NAME -n OPERATOR_NAMESPACE -c nvidia-driver-ctr -- nvidia-smi

Fri Sep 25 16:20:15 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 595.91.07              Driver Version: 595.91.07      CUDA Version: 13.2     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5080        On  |   00000000:00:06.0 Off |                  N/A |
|  0%   39C    P8             16W /  360W |       1MiB /  16303MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
|   1  NVIDIA GeForce RTX 5080        On  |   00000000:00:07.0 Off |                  N/A |
|  0%   38C    P8             11W /  360W |       1MiB /  16303MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
|   2  NVIDIA GeForce RTX 5080        On  |   00000000:00:08.0 Off |                  N/A |
|  0%   39C    P8             15W /  360W |       1MiB /  16303MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
|   3  NVIDIA GeForce RTX 5080        On  |   00000000:00:09.0 Off |                  N/A |
|  0%   38C    P8             12W /  360W |       1MiB /  16303MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

Details
Using AI assistant to debug, I managed to find the following cause:

GPU Operator v26.7.1 defines:

  # assets/state-mps-control-daemon/0400_daemonset.yaml
  command: [mps-control-daemon]

and:

  - name: SEND_SIGNAL
    value: "true"
  - name: SIGNAL
    value: "1"
  - name: PROCESS_TO_SIGNAL
    value: "/usr/bin/mps-control-daemon"

Sources:

config-manager uses an exact command-line match

In k8s-device-plugin v0.20.1, config-manager searches processes and compares the first command-line element exactly:

  if cmdline[0] == f.ProcessToSignal {
      return p.PID, nil
  }

Source:

https://github.com/NVIDIA/k8s-device-plugin/blob/v0.20.1/cmd/config-manager/main.go#L424-L450

The same file returns the observed error if no matching process is found:

  return fmt.Errorf("error finding pid: %v", err)

Source:

https://github.com/NVIDIA/k8s-device-plugin/blob/v0.20.1/cmd/config-manager/main.go#L424-L433

Dominant language: Go
Stars: 2.9k
Forks: 552
Avg merge: 1d 21h
Merged PRs (30d): 78

Getting set up

Read the contributing guide

First steps

Read the whole issue, then the project's contributing guide.
Comment on the issue to say you are picking it up — it saves two people doing the same work.
Fork the repository and make your change on a branch.
Open a pull request that references the issue number.

More from NVIDIA/gpu-operator

DaemonSet readiness accepts status from an older generation
Open
- Is something wrong?
Difficulty 2/5 1-3 hours Newbie friendliness 84/100

NVIDIA/gpu-operator#2968 · Opened Sep 25, 2026

Maintainers usually reply within 1 day
[Bug]: GPUCluster common name label breaks DRA validator selector
Open
- Is something wrong?
Difficulty 2/5 1-3 hours Newbie friendliness 88/100

NVIDIA/gpu-operator#2955 · Opened Sep 23, 2026

Maintainers usually reply within 1 day
[Question]: Is the omission of env support for the operator Deployment intentional?
Open
- Is something wrong?
lifecycle/stale question

Difficulty 2/5 1-3 hours Newbie friendliness 64/100

NVIDIA/gpu-operator#2280 · 2 comments · 1 reaction · Opened Apr 9, 2026

Maintainers usually reply within 1 day
[Bug]: GPUCluster ignores daemonsets.updateStrategy=OnDelete
Open
- Is something wrong?
Difficulty 4/5 3-5 days Newbie friendliness 68/100

NVIDIA/gpu-operator#2957 · Opened Sep 23, 2026

Maintainers usually reply within 1 day
[Feature]: add metricRelabelings field to the dcgm exporter service monitor.
Open
- Is something wrong?
feature lifecycle/frozen needs-triage

Difficulty 3/5 1-2 days Newbie friendliness 65/100

NVIDIA/gpu-operator#2938 · Opened Sep 21, 2026

Maintainers usually reply within 1 day

All issues in NVIDIA/gpu-operator

Similar issues

`aidy pr` generates PR bodies with `##` section headers
Open
- Is something wrong?
enhancement

Difficulty 2/5 1-3 hours Newbie friendliness 75/100

volodya-lombrozo/aidy#333 · Opened Sep 26, 2026
test: TestStartProxyTakesPresetAndIdleExitAsArguments races its own fake proxy and fails under CI load
Open
- Is something wrong?
Difficulty 2/5 1-3 hours Newbie friendliness 85/100

rossoctl/context-guru#314 · Opened Sep 26, 2026

Maintainers usually reply within 1 day
429 from oneshot-free: pinned iosAppVersion 26.42 is stale (DeepL iOS is 26.52)
Open
- Is something wrong?
Difficulty 1/5 Under an hour Newbie friendliness 86/100

OwO-Network/DLX#236 · 1 comment · 1 reaction · Opened Sep 26, 2026
TRIM leaves duplicate spaces between words
Open
- Is something wrong?
Difficulty 2/5 1-3 hours Newbie friendliness 78/100

qax-os/excelize#2430 · Opened Sep 26, 2026

Maintainers usually reply within 1 day
flagd-proxy Deployment fails to create when flagdProxyConfiguration.image.tag is digest-pinned (tag@sha256:...)
Open
- Is something wrong?
Difficulty 2/5 1-3 hours Newbie friendliness 84/100

open-feature/open-feature-operator#863 · Opened Sep 26, 2026

More Go issues

[Bug]: GPU Operator MPS config-manager cannot signal MPS daemon due to process-target mismatch

Is something wrong?

Assessment

Research direction

Description

Getting set up

First steps

More from NVIDIA/gpu-operator

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Similar issues

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Get new issues in your inbox