add linux nodeselector

[timmy-wright]

While running Reloader in my cluster, the pod was deployed to a windows node and failed as the image doesn't support windows.

This PR adds a node selector to the pod so it will only be scheduled on Linux nodes.

Related to #368

Pods

tim@wrightt-laptop:~/git/Reloader$ kubectl get pods -o wide
NAME                                 READY   STATUS         RESTARTS   AGE   IP             NODE              NOMINATED NODE   READINESS GATES
hpc-2019-85jsr                       1/1     Running        0          31m   10.224.0.5     aksws2019000000   <none>           <none>
hpc-2022-vxnpt                       1/1     Running        0          33m   10.224.0.10    aksws2022000003   <none>           <none>
hpc-2025-2022-cbdd8                  1/1     Running        0          33m   10.224.0.11    aksws2025000000   <none>           <none>
hpc-2025-dtkv9                       1/1     Running        0          31m   10.224.0.11    aksws2025000000   <none>           <none>
hpc-23h2-ldphj                       1/1     Running        0          33m   10.224.0.7     akswsannu000000   <none>           <none>
reloader-reloader-5877c96d8f-wfb79   0/1     ErrImagePull   0          11m   172.16.0.181   aksws2019000000   <none>           <none>
windows-updates-6g9n6                1/1     Running        0          14m   10.224.0.11    aksws2025000000   <none>           <none>
windows-updates-hr9jr                1/1     Running        0          14m   10.224.0.7     akswsannu000000   <none>           <none>
windows-updates-kxjsq                1/1     Running        0          14m   10.224.0.5     aksws2019000000   <none>           <none>
windows-updates-lfbmv                1/1     Running        0          14m   10.224.0.10    aksws2022000003   <none>           <none>

Logs

tim@wrightt-laptop:~/git/Reloader$ kubectl logs reloader-reloader-5877c96d8f-wfb79
Error from server (BadRequest): container "reloader-reloader" in pod "reloader-reloader-5877c96d8f-wfb79" is waiting to start: trying and failing to pull image

Describe Container

tim@wrightt-laptop:~/git/Reloader$ kubectl describe pod reloader-reloader-5877c96d8f-wfb79
Name:             reloader-reloader-5877c96d8f-wfb79
Namespace:        default
Priority:         0
Service Account:  reloader-reloader
Node:             aksws2019000000/10.224.0.5
Start Time:       Sat, 08 Nov 2025 10:10:39 +1300
Labels:           app=reloader-reloader
                  app.kubernetes.io/instance=reloader
                  app.kubernetes.io/managed-by=Helm
                  app.kubernetes.io/name=reloader
                  app.kubernetes.io/version=v1.4.10
                  chart=reloader-2.2.5
                  group=com.stakater.platform
                  helm.sh/chart=reloader-2.2.5
                  heritage=Helm
                  pod-template-hash=5877c96d8f
                  provider=stakater
                  release=reloader
                  version=v1.4.10
Annotations:      <none>
Status:           Pending
SeccompProfile:   RuntimeDefault
IP:               172.16.0.181
IPs:
  IP:           172.16.0.181
Controlled By:  ReplicaSet/reloader-reloader-5877c96d8f
Containers:
  reloader-reloader:
    Container ID:
    Image:         ghcr.io/stakater/reloader:v1.4.10
    Image ID:
    Port:          9090/TCP
    Host Port:     0/TCP
    Args:
      --log-level=info
    State:          Waiting
      Reason:       ImagePullBackOff
    Ready:          False
    Restart Count:  0
    Liveness:       http-get http://:http/live delay=10s timeout=5s period=10s #success=1 #failure=5
    Readiness:      http-get http://:http/metrics delay=10s timeout=5s period=10s #success=1 #failure=5
    Environment:
      GOMAXPROCS:                node allocatable (limits.cpu)
      GOMEMLIMIT:                node allocatable (limits.memory)
      RELOADER_NAMESPACE:        default (v1:metadata.namespace)
      RELOADER_DEPLOYMENT_NAME:  reloader-reloader
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-7j6kg (ro)
Conditions:
  Type                        Status
  PodReadyToStartContainers   True
  Initialized                 True
  Ready                       False
  ContainersReady             False
  PodScheduled                True
Volumes:
  kube-api-access-7j6kg:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   BestEffort
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  10m                  default-scheduler  Successfully assigned default/reloader-reloader-5877c96d8f-wfb79 to aksws2019000000
  Normal   Pulling    7m21s (x5 over 10m)  kubelet            Pulling image "ghcr.io/stakater/reloader:v1.4.10"
  Warning  Failed     7m20s (x5 over 10m)  kubelet            Failed to pull image "ghcr.io/stakater/reloader:v1.4.10": rpc error: code = NotFound desc = failed to pull and unpack image "ghcr.io/stakater/reloader:v1.4.10": no match for platform in manifest: not found
  Warning  Failed     7m20s (x5 over 10m)  kubelet            Error: ErrImagePull
  Normal   BackOff    15s (x42 over 10m)   kubelet            Error: ImagePullBackOff

[msafwankarim]

Instead of enforcing it in the chart, we should add this in values.yaml under reloader.deployment.nodeSelector by default. This will allow the user to override or unset it if they don't need it for any reason

[msafwankarim]

Currently it is setting the OS nodeSelector only if the nodeSelector is already provided, that's why test cases are failing. Adding the nodeSelector in values file instead of deployment.yaml will work