installing flux view an on demand volume is almost working!

docs/_static/data/addons.json

@@ -43,5 +43,10 @@
   "name": "volume-secret",
   "description": "secret volume type",
   "family": "volume"
+ },
+ {
+  "name": "workload-flux",
+  "description": "hierarchical graph-based scheduler and resource manager",
+  "family": "workload"
  }
 ]

docs/getting_started/addons.md

@@ -157,6 +157,63 @@ spec:
 
 **Note that we have support for a custom application container, but haven't written any good examples yet!**
 
+## Workload
+
+### workload-flux
+
+If you need to "throw in" Flux Framework into your container to use as a scheduler, you can do that with an addon!
+
+> Yes, it's astounding. 🦩️
+
+This works by way of the same trick that we use for other addons that have a complex (and/or large) install setup. We:
+
+- Build the software into an isolated spack "copy" view
+- The software is then (generally) at some `/opt/view` and `/opt/software`
+- The flux container is added as a sidecar container to your pod for your replicated job
+  - Additional setup / configuration is done here
+- We can then create an empty volume that is shared by your metric or scaled application
+- The entire tree is copied over into the empty volume
+- When the copy is done, indicated by the final touch of a file, the updated container entrypoint is run
+- This typically means we have taken your metric command, and wrapped it in a Flux submit.
+
+It's really cool because it means you can run a metric / application with Flux without needing
+to install it into your container to begin with. The one important detail is a matching of
+general operating system. The current view uses rocky, however the image is customizable
+(and we can provide other bases if/when requested). Here are the arguments you can customize
+under the metric -> options.
+
+| Name | Description | Type | Default |
+|-----|-------------|------------|------|
+| mount | Path to mount flux view in application container | string | /opt/share |
+| tasks | Number of tasks `-n` to give to flux (not provided if not set) | string | unset |
+| image | Customize the container image | string | `ghcr.io/rse-ops/spack-flux-rocky-view:tag-8` |
+| fluxUser  | The flux user (currently not used, but TBA)  | string | flux |
+| fluxUid  | The flux user ID (currently not used, but TBA)  | string | 1004 |
+| interactive  | Run flux in interactive mode  | string | "false" |
+| connectTimeout | How long zeroMQ should wait to retry | string | "5s" |
+| quorum | The number of brokers to require before starting the cluster | string | (total brokers or pods) |
+| debugZeroMQ | Turn on zeroMQ debugging | string | "false" |
+| logLevel | Customize the flux log level | string | "6" |
+| queuePolicy | Queue policy for flux to use | string | fcfs |
+| workerLetter | The letter that the worker job is expected to have | string | w |
+| launcherLetter | The letter that the launcher job is expected to have | string | w |
+| workerIndex | The index of the replicated job for the worker | string | 0 |
+| launcherIndex | The index of the replicated job for the launcher | string | 0 |
+| preCommand | Pre-command logic to run in launcher/workers before flux is started (after setup in flux container) | string | unset |
+
+Note that the number of pods for flux defaults to the number in your MetricSet, along 
+with the namespace and service name.
+
+**Important** the flux addon is currently supported for metric types that:
+
+1. have the launcher / worker design (so the hostlist.txt is present in the PWD)
+2. Have scp installed, as the shared certificate needs to be copied from the lead broker to all followers
+3. Ideally have munge installed - we do try to install it (but better to already be there)
+
+We also currently run flux as root. This is considered bad practice, but probably OK
+for this early development work. We don't see a need to have shared namespace / operator
+environments at this point, which is why I didn't add it.
+
 ## Performance
 
 ### perf-hpctoolkit

examples/addons/flux-lammps/metrics.yaml

@@ -0,0 +1,32 @@
+apiVersion: flux-framework.org/v1alpha2
+kind: MetricSet
+metadata:
+  labels:
+    app.kubernetes.io/name: metricset
+    app.kubernetes.io/instance: metricset-sample
+  name: metricset-sample
+spec:
+  # Number of pods for lammps (one launcher, the rest workers)
+  pods: 4
+  logging:
+    interactive: true
+
+  metrics:
+
+   # Running more scaled lammps is our main goal
+   - name: app-lammps
+
+     # This flux addon is built on rocky, and we can provide additional os bases
+     image: ghcr.io/converged-computing/metric-lammps-intel-mpi:rocky
+
+     options:
+       command: lmp -v x 2 -v y 2 -v z 2 -in in.reaxc.hns -nocite
+       workdir: /opt/lammps/examples/reaxff/HNS
+
+     # Add on hpctoolkit, will mount a volume and wrap lammps
+     addons:
+       - name: workload-flux
+         options: 
+           # Ensure intel environment is setup
+           preCommand: . /opt/intel/mpi/latest/env/vars.sh
+           workdir: /opt/lammps/examples/reaxff/HNS

pkg/addons/addons.go

@@ -26,6 +26,7 @@ var (
 	AddonFamilyPerformance = "performance"
 	AddonFamilyVolume      = "volume"
 	AddonFamilyApplication = "application"
+	AddonFamilyWorkload    = "workload"
 )
 
 // A general metric is a container added to a JobSet
@@ -37,7 +38,7 @@ type Addon interface {
 	Description() string
 
 	// Options and exportable attributes
-	SetOptions(*api.MetricAddon)
+	SetOptions(*api.MetricAddon, *api.MetricSet)
 	Options() map[string]intstr.IntOrString
 	ListOptions() map[string][]intstr.IntOrString
 	MapOptions() map[string]map[string]intstr.IntOrString
@@ -65,7 +66,7 @@ type AddonBase struct {
 	mapOptions  map[string]map[string]intstr.IntOrString
 }
 
-func (b *AddonBase) SetOptions(metric *api.MetricAddon)                                   {}
+func (b *AddonBase) SetOptions(addon *api.MetricAddon, metric *api.MetricSet)             {}
 func (b *AddonBase) CustomizeEntrypoints([]*specs.ContainerSpec, []*jobset.ReplicatedJob) {}
 
 func (b *AddonBase) Validate() bool {
@@ -97,7 +98,7 @@ func (b *AddonBase) MapOptions() map[string]map[string]intstr.IntOrString {
 }
 
 // GetAddon looks up and validates an addon
-func GetAddon(a *api.MetricAddon) (Addon, error) {
+func GetAddon(a *api.MetricAddon, set *api.MetricSet) (Addon, error) {
 
 	// We don't want to change the addon interface/struct itself
 	template, ok := Registry[a.Name]
@@ -111,7 +112,7 @@ func GetAddon(a *api.MetricAddon) (Addon, error) {
 	addon := reflect.New(templateType.Type()).Interface().(Addon)
 
 	// Set options before validation
-	addon.SetOptions(a)
+	addon.SetOptions(a, set)
 
 	// Validate the addon
 	if !addon.Validate() {

pkg/addons/commands.go

@@ -42,9 +42,9 @@ func (a *PerfAddon) CustomizeEntrypoints(
 	}
 }
 
-func (a *PerfAddon) SetOptions(metric *api.MetricAddon) {
+func (a *PerfAddon) SetOptions(addon *api.MetricAddon, metric *api.MetricSet) {
 	a.Identifier = perfCommandsName
-	a.SetSharedCommandOptions(metric)
+	a.SetSharedCommandOptions(addon)
 }
 
 // addContainerCaps adds capabilities to a container spec
@@ -102,9 +102,9 @@ func (m CommandAddon) Family() string {
 	return AddonFamilyApplication
 }
 
-func (a *CommandAddon) SetOptions(metric *api.MetricAddon) {
+func (a *CommandAddon) SetOptions(addon *api.MetricAddon, metric *api.MetricSet) {
 	a.Identifier = commandsName
-	a.SetSharedCommandOptions(metric)
+	a.SetSharedCommandOptions(addon)
 }
 
 // Set custom options / attributes for the metric

pkg/addons/containers.go

@@ -139,8 +139,8 @@ func (a *ApplicationAddon) setDefaultEntrypoint() {
 }
 
 // Calling the default allows a custom application that uses this to do the same
-func (a *ApplicationAddon) SetOptions(metric *api.MetricAddon) {
-	a.SetDefaultOptions(metric)
+func (a *ApplicationAddon) SetOptions(addon *api.MetricAddon, metric *api.MetricSet) {
+	a.SetDefaultOptions(addon)
 }
 
 // Underlying function that can be shared