chart(jobset): add controller.hostNetwork toggle (#978)

* chart(jobset): add controller.hostNetwork toggle

Signed-off-by: Mohammad Mamdouhi <m.mamdouhi@instadeep.com>

* charts(jobset): document controller.hostNetwork and regen README

Signed-off-by: Mohammad Mamdouhi <m.mamdouhi@instadeep.com>

---------

Signed-off-by: Mohammad Mamdouhi <m.mamdouhi@instadeep.com>

Author: m-mamdouhi

charts/jobset/README.md

@@ -88,6 +88,7 @@ See [helm uninstall](https://helm.sh/docs/helm/helm_uninstall) for command docum
 | controller.affinity | object | `{}` | Affinity of the jobset controller pods. |
 | controller.tolerations | list | `[]` | Tolerations of the jobset controller pods. |
 | controller.podSecurityContext | object | `{"runAsNonRoot":true,"seccompProfile":{"type":"RuntimeDefault"}}` | Security context of all jobset controller containers. |
+| controller.hostNetwork | bool | `false` | Run the controller/webhook Pods on the node’s network namespace instead of the overlay CNI. |
 | certManager.enable | bool | `false` | Whether to use cert-manager to generate certificates for the jobset webhook. |
 | certManager.issuerRef | object | `{}` | The reference to the issuer. If empty, self-signed issuer will be created and used. |
 | prometheus.enable | bool | `false` | Whether to enable Prometheus metrics exporting. |

charts/jobset/templates/controller/deployment.yaml

@@ -137,3 +137,7 @@ spec:
         {{- toYaml . | nindent 8 }}
       {{- end }}
       serviceAccountName: {{ include "jobset.controller.serviceAccount.name" . }}
+      {{- if .Values.controller.hostNetwork }}
+      hostNetwork: true
+      dnsPolicy: ClusterFirstWithHostNet
+      {{- end }}

charts/jobset/values.yaml

@@ -75,6 +75,8 @@ controller:
     runAsNonRoot: true
     seccompProfile:
       type: RuntimeDefault
+  # -- Run the controller/webhook Pods on the node’s network namespace instead of the overlay CNI.
+  hostNetwork: false
 certManager:
   # -- Whether to use cert-manager to generate certificates for the jobset webhook.
   enable: false

site/content/en/docs/reference/jobset.v1alpha2.md

@@ -246,6 +246,40 @@ An empty list will apply to all replicatedJobs.</p>
 </tbody>
 </table>
 
+## `IndividualJobStatus`     {#jobset-x-k8s-io-v1alpha2-IndividualJobStatus}
+    
+
+**Appears in:**
+
+- [JobSetStatus](#jobset-x-k8s-io-v1alpha2-JobSetStatus)
+
+
+<p>IndividualJobStatus holds the status of an individual Job within a ReplicatedJob.</p>
+
+
+<table class="table">
+<thead><tr><th width="30%">Field</th><th>Description</th></tr></thead>
+<tbody>
+    
+  
+<tr><td><code>name</code> <B>[Required]</B><br/>
+<code>string</code>
+</td>
+<td>
+   <p>Name of the Job.</p>
+</td>
+</tr>
+<tr><td><code>recreates</code> <B>[Required]</B><br/>
+<code>int32</code>
+</td>
+<td>
+   <p>Recreates is the number of times an individual Job has been recreated.
+This counter is reset to 0 if the parent ReplicatedJob or JobSet is restarted.</p>
+</td>
+</tr>
+</tbody>
+</table>
+
 ## `JobSetRestartStrategy`     {#jobset-x-k8s-io-v1alpha2-JobSetRestartStrategy}
 
 (Alias of `string`)
@@ -417,12 +451,11 @@ It can be either Completed or Failed. Otherwise, it is empty by default.</p>
    <p>ReplicatedJobsStatus track the number of JobsReady for each replicatedJob.</p>
 </td>
 </tr>
-<tr><td><code>individualJobRecreates</code><br/>
-<code>map[string]int32</code>
+<tr><td><code>individualJobsStatus</code><br/>
+<a href="#jobset-x-k8s-io-v1alpha2-IndividualJobStatus"><code>[]IndividualJobStatus</code></a>
 </td>
 <td>
-   <p>IndividualJobRecreates tracks the number of times an individual Job within
-the JobSet has been recreated (i.e. in case of RecreateJob failure policy).</p>
+   <p>IndividualJobsStatus tracks the status of individual Jobs within ReplicatedJobs.</p>
 </td>
 </tr>
 </tbody>