podspec: add additional features for podspec

Dockerfile

@@ -1,5 +1,5 @@
 # Build the manager binary
-FROM golang:1.20 as builder
+FROM golang:1.20 AS builder
 
 WORKDIR /workspace
 

api/v1alpha2/minicluster_types.go

@@ -188,6 +188,10 @@ type PodSpec struct {
 	// +optional
 	ServiceAccountName string `json:"serviceAccountName,omitempty"`
 
+	// RuntimeClassName for the pod
+	// +optional
+	RuntimeClassName string `json:"runtimeClassName,omitempty"`
+
 	// Automatically mount the service account name
 	// +optional
 	AutomountServiceAccountToken bool `json:"automountServiceAccountToken,omitempty"`

api/v1alpha2/swagger.json

@@ -775,6 +775,10 @@
           "description": "Restart Policy",
           "type": "string"
         },
+        "runtimeClassName": {
+          "description": "RuntimeClassName for the pod",
+          "type": "string"
+        },
         "schedulerName": {
           "description": "Scheduler name for the pod",
           "type": "string"

chart/templates/minicluster-crd.yaml

@@ -555,6 +555,9 @@ spec:
                   restartPolicy:
                     description: Restart Policy
                     type: string
+                  runtimeClassName:
+                    description: RuntimeClassName for the pod
+                    type: string
                   schedulerName:
                     description: Scheduler name for the pod
                     type: string

config/crd/bases/flux-framework.org_miniclusters.yaml

@@ -558,6 +558,9 @@ spec:
                   restartPolicy:
                     description: Restart Policy
                     type: string
+                  runtimeClassName:
+                    description: RuntimeClassName for the pod
+                    type: string
                   schedulerName:
                     description: Scheduler name for the pod
                     type: string

controllers/flux/job.go

@@ -45,8 +45,8 @@ func NewMiniClusterJob(cluster *api.MiniCluster) (*batchv1.Job, error) {
 			Labels:    cluster.Spec.JobLabels,
 		},
 
+		// Completions must be == to Parallelism to allow for scaling
 		Spec: batchv1.JobSpec{
-
 			BackoffLimit:          &backoffLimit,
 			Completions:           &cluster.Spec.Size,
 			Parallelism:           &cluster.Spec.Size,
@@ -70,13 +70,22 @@ func NewMiniClusterJob(cluster *api.MiniCluster) (*batchv1.Job, error) {
 					ImagePullSecrets:             getImagePullSecrets(cluster),
 					ServiceAccountName:           cluster.Spec.Pod.ServiceAccountName,
 					AutomountServiceAccountToken: &cluster.Spec.Pod.AutomountServiceAccountToken,
-					RestartPolicy:                corev1.RestartPolicy(cluster.Spec.Pod.RestartPolicy),
+					RestartPolicy:                corev1.RestartPolicyOnFailure,
 					NodeSelector:                 cluster.Spec.Pod.NodeSelector,
 					SchedulerName:                cluster.Spec.Pod.SchedulerName,
 				},
 			},
 		},
 	}
+	// Custom restart policy
+	if cluster.Spec.Pod.RestartPolicy != "" {
+		job.Spec.Template.Spec.RestartPolicy = corev1.RestartPolicy(cluster.Spec.Pod.RestartPolicy)
+	}
+
+	// Only add runClassName if defined
+	if cluster.Spec.Pod.RuntimeClassName != "" {
+		job.Spec.Template.Spec.RuntimeClassName = &cluster.Spec.Pod.RuntimeClassName
+	}
 
 	// Add Affinity to map one pod / node only if the user hasn't disbaled it
 	if !cluster.Spec.Network.DisableAffinity {

controllers/flux/pods.go

@@ -135,13 +135,23 @@ func (r *MiniClusterReconciler) newServicePod(
 			SetHostnameAsFQDN:            &setAsFQDN,
 			Volumes:                      existingVolumes,
 			ImagePullSecrets:             getImagePullSecrets(cluster),
-			RestartPolicy:                corev1.RestartPolicy(cluster.Spec.Pod.RestartPolicy),
+			RestartPolicy:                corev1.RestartPolicyOnFailure,
 			ServiceAccountName:           cluster.Spec.Pod.ServiceAccountName,
 			AutomountServiceAccountToken: &cluster.Spec.Pod.AutomountServiceAccountToken,
 			NodeSelector:                 cluster.Spec.Pod.NodeSelector,
 		},
 	}
 
+	// Custom restart policy
+	if cluster.Spec.Pod.RestartPolicy != "" {
+		pod.Spec.RestartPolicy = corev1.RestartPolicy(cluster.Spec.Pod.RestartPolicy)
+	}
+
+	// Only add runClassName if defined
+	if cluster.Spec.Pod.RuntimeClassName != "" {
+		pod.Spec.RuntimeClassName = &cluster.Spec.Pod.RuntimeClassName
+	}
+
 	// Assemble existing volume mounts - they are added with getContainers
 	mounts := []corev1.VolumeMount{}
 

docs/getting_started/custom-resource-definition.md

@@ -188,8 +188,11 @@ When enabled, meaning that we use flux from a view within the container, these c
 
  - [ghcr.io/converged-computing/flux-view-rocky:tag-9](https://github.com/converged-computing/flux-views/pkgs/container/flux-view-rocky)
  - [ghcr.io/converged-computing/flux-view-rocky:tag-8](https://github.com/converged-computing/flux-views/pkgs/container/flux-view-rocky)
+ - [ghcr.io/converged-computing/flux-view-ubuntu:tag-noble](https://github.com/converged-computing/flux-views/pkgs/container/flux-view-ubuntu)
+ - [ghcr.io/converged-computing/flux-view-ubuntu:tag-jammy](https://github.com/converged-computing/flux-views/pkgs/container/flux-view-ubuntu)
  - [ghcr.io/converged-computing/flux-view-ubuntu:tag-focal](https://github.com/converged-computing/flux-views/pkgs/container/flux-view-ubuntu)
 
+
 Note that we have [arm builds](https://github.com/converged-computing/flux-views/tree/main/arm) available for each of rocky and ubuntu as well.
 If you don't want to use Flux from a view (and want to use the v1apha1 design of the Flux Operator that had the application alongside Flux) you can do that by way of disabling the flux view:
 
@@ -682,6 +685,34 @@ pod:
   serviceAccountName: my-service-account
 ```
 
+#### restartPolicy
+
+To customize the restartPolicy for the pod:
+
+```yaml
+pod:
+  restartPolicy: Never
+```
+
+#### runtimeClassName
+
+To add a runtime class name:
+
+```yaml
+pod:
+  runtimeClassName: nvidia
+```
+
+#### automountServiceAccountToken
+
+If you want to automatically mount a service account token:
+
+```yaml
+pod:
+  automountServiceAccountToken: true
+```
+
+
 #### nodeSelector
 
 A node selector is a set of key value pairs that helps to schedule pods to the right nodes! You can
@@ -720,10 +751,8 @@ name: rabbit
 
 #### image
 
-This is the only required attribute! You *must* provide a container base that has Flux.
-The requirements of your container are defined in the README of the [flux-hpc](https://github.com/rse-ops/flux-hpc/)
-repository. Generally speaking, you need to have Flux executables, Flux Python bindings,
-and your own executables on the path, and should be started with root with a flux user.
+You do not need to provide a container base that has Flux, but you must make sure your view (with a particular operator system) that will add Flux matches your container. We don't require you to start as root, but if you
+have a container with a non-root user, the user needs to have sudo available (to act as root).
 If you use the [fluxrm/flux-sched](https://hub.docker.com/r/fluxrm/flux-sched)
 base containers this is usually a good start.
 

examples/dist/flux-operator-arm.yaml

@@ -564,6 +564,9 @@ spec:
                   restartPolicy:
                     description: Restart Policy
                     type: string
+                  runtimeClassName:
+                    description: RuntimeClassName for the pod
+                    type: string
                   schedulerName:
                     description: Scheduler name for the pod
                     type: string

examples/dist/flux-operator.yaml

@@ -564,6 +564,9 @@ spec:
                   restartPolicy:
                     description: Restart Policy
                     type: string
+                  runtimeClassName:
+                    description: RuntimeClassName for the pod
+                    type: string
                   schedulerName:
                     description: Scheduler name for the pod
                     type: string

pkg/flux/templates/components.sh

@@ -21,7 +21,7 @@ url=$goshareUrl/wait-fs-{{ .Spec.Flux.Arch }}
 # This waiting script is intended to wait for the flux view, and then start running
 curl -L -O -s -o ./wait-fs -s ${url} {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }} || wget ${url} -q -O ./wait-fs {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }} || true
 chmod +x ./wait-fs || true
-mv ./wait-fs /usr/bin/goshare-wait-fs || true
+${SUDO} mv ./wait-fs /usr/bin/goshare-wait-fs || true
 
 # Ensure spack view is on the path, wherever it is mounted
 viewbase="{{ .ViewBase }}"
@@ -47,9 +47,9 @@ goshare-wait-fs -p ${viewbase}/flux-operator-done.txt {{ if .Spec.Logging.Quiet
 # Copy mount software to /opt/software
 # If /opt/software already exists, we need to copy into it
 if [[ -e  "/opt/software" ]]; then
-  cp -R ${viewbase}/software/* /opt/software/ || true
+  ${SUDO} cp -R ${viewbase}/software/* /opt/software/ || true
 else
-  cp -R ${viewbase}/software /opt/software || true
+  ${SUDO} cp -R ${viewbase}/software /opt/software || true
 fi
 {{end}}
 
@@ -72,10 +72,10 @@ echo "Python root: $foundroot" {{ if .Spec.Logging.Quiet }} > /dev/null 2>&1{{ e
 
 # If we found the right python, ensure it's linked (old link does not work)
 if [[ -f "${pythonversion}" ]]; then
-   rm -rf $viewroot/bin/python3
-   rm -rf $viewroot/bin/python
-   ln -s ${pythonversion} $viewroot/lib/python  || true
-   ln -s ${pythonversion} $viewroot/lib/python3 || true
+   ${SUDO} rm -rf $viewroot/bin/python3
+   ${SUDO} rm -rf $viewroot/bin/python
+   ${SUDO} ln -s ${pythonversion} $viewroot/lib/python  || true
+   ${SUDO} ln -s ${pythonversion} $viewroot/lib/python3 || true
 fi
 
 # Ensure we use flux's python (TODO update this to use variable)
@@ -87,15 +87,16 @@ find $viewroot . -name libpython*.so* {{ if .Spec.Logging.Quiet }}> /dev/null 2>
 ls -l /mnt/flux/view/lib/libpython3.11.so.1.0 {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }}
 
 # Write an easy file we can source for the environment
-cat <<EOT >> ${viewbase}/flux-view.sh
+cat <<EOT >> ./flux-view.sh
 #!/bin/bash
 export PATH=$PATH
 export PYTHONPATH=$PYTHONPATH
 export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:$viewroot/lib
 export fluxsocket=local://${viewroot}/run/flux/local
 EOT
+${SUDO} mv ./flux-view.sh ${viewbase}/flux-view.sh
 {{end}}
 {{define "ensure-pip"}}
-${pythonversion} -m pip --version || ${pythonversion} -m ensurepip || (wget https://bootstrap.pypa.io/get-pip.py && ${pythonversion} ./get-pip.py) {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }}
-${pythonversion} -m pip --upgrade pip {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }}
-{{end}}
+${SUDO} ${pythonversion} -m pip --version || ${SUDO} ${pythonversion} -m ensurepip || (${SUDO} wget https://bootstrap.pypa.io/get-pip.py && ${pythonversion} ./get-pip.py) {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }}
+${SUDO} ${pythonversion} -m pip --upgrade pip {{ if .Spec.Logging.Quiet }}> /dev/null 2>&1{{ end }}
+{{end}}

pkg/flux/templates/wait.sh

@@ -6,6 +6,18 @@
 # We use the actual time command and not the wrapper, otherwise we get there is no argument -f
 {{ if .Spec.Logging.Timed }}which /usr/bin/time > /dev/null 2>&1 || (echo "/usr/bin/time is required to use logging.timed true" && exit 1);{{ end }}
 
+# Set the flux user and id from the getgo
+fluxuser=$(whoami)
+fluxuid=$(id -u $fluxuser)
+
+# Add fluxuser to sudoers living... dangerously!
+# A non root user container requires sudo to work
+SUDO=""
+if [[ "${fluxuser}" != "root" ]]; then
+  echo "${fluxuser} ALL=(ALL) NOPASSWD: ALL" >> /etc/sudoers
+  SUDO="sudo"
+fi
+
 # If any initCommand logic is defined
 {{ .Container.Commands.Init}} {{ if .Spec.Logging.Quiet }}> /dev/null{{ end }}
 
@@ -14,10 +26,6 @@
 {{template "wait-view" .}}
 {{ if not .Spec.Flux.Container.Disable }}{{template "paths" .}}{{ end }}
 
-# Set the flux user and id from the getgo
-fluxuser=$(whoami)
-fluxuid=$(id -u $fluxuser)
-
 # Variables we can use again
 cfg="${viewroot}/etc/flux/config"
 command="{{ .Container.Command }}"
@@ -28,19 +36,21 @@ command="{{ .Container.Command }}"
 {{ if not .Spec.Logging.Quiet }}
 echo 
 echo "Hello user ${fluxuser}"{{ end }}
-
-# Add fluxuser to sudoers living... dangerously!
-if [[ "${fluxuser}" != "root" ]]; then
-  echo "${fluxuser} ALL=(ALL) NOPASSWD: ALL" >> /etc/sudoers
-fi
 
 # Ensure the flux user owns the curve.cert
 # We need to move the curve.cert because config map volume is read only
 curvesrc=/flux_operator/curve.cert
 curvepath=$viewroot/curve/curve.cert
 
-mkdir -p $viewroot/curve
-cp $curvesrc $curvepath
+# run directory must be owned by this user
+# and /var/lib/flux
+if [[ "${fluxuser}" != "root" ]]; then
+  ${SUDO} chown -R ${fluxuser} ${viewroot}/run/flux ${viewroot}/var/lib/flux
+fi
+
+# Prepare curve certificate!
+${SUDO} mkdir -p $viewroot/curve
+${SUDO} cp $curvesrc $curvepath
 {{ if not .Spec.Logging.Quiet }}
 echo 
 echo "🌟️ Curve Certificate"
@@ -49,9 +59,9 @@ cat ${curvepath}
 {{ end }}
 
 # Remove group and other read
-chmod o-r ${curvepath}
-chmod g-r ${curvepath}
-chown -R ${fluxuid} ${curvepath}
+${SUDO} chmod o-r ${curvepath}
+${SUDO} chmod g-r ${curvepath}
+${SUDO} chown -R ${fluxuid} ${curvepath}
 
 # If we have disabled the view, we need to use the flux here to generate resources
 {{ if .Spec.Flux.Container.Disable }}
@@ -61,7 +71,8 @@ echo
 echo "📦 Resources"
 echo "flux R encode --hosts=${hosts} --local"
 {{ end }}
-flux R encode --hosts=${hosts} --local > ${viewroot}/etc/flux/system/R
+flux R encode --hosts=${hosts} --local > /tmp/R
+${SUDO} mv /tmp/R ${viewroot}/etc/flux/system/R
 {{ if not .Spec.Logging.Quiet }}cat ${viewroot}/etc/flux/system/R{{ end }}
 {{ end }}
 

sdk/python/v1alpha2/docs/PodSpec.md

@@ -11,6 +11,7 @@ Name | Type | Description | Notes
 **node_selector** | **dict[str, str]** | NodeSelectors for a pod | [optional] 
 **resources** | [**dict[str, IntOrString]**](IntOrString.md) | Resources include limits and requests | [optional] 
 **restart_policy** | **str** | Restart Policy | [optional] 
+**runtime_class_name** | **str** | RuntimeClassName for the pod | [optional] 
 **scheduler_name** | **str** | Scheduler name for the pod | [optional] 
 **service_account_name** | **str** | Service account name for the pod | [optional]