Add fqdn

[AlterHoodie]

Why are these changes needed?

Ray worker Pods are only addressable by Pod IP today, and that IP changes whenever a Pod is recreated. This PR sets worker hostname / subdomain from replica (and host - for multihost) indices and always creates the RayCluster headless Service so each worker gets a stable Pod FQDN that can survive recreates.

Hostname/subdomain assignment still sits behind RayMultiHostIndexing, since that is where replica indices are assigned. Creating the headless Service is no longer limited to multi-host groups (NumOfHosts > 1).

That stretches the meaning of RayMultiHostIndexing. Looking for guidance on whether to:

1. Add a separate feature flag for worker FQDNs / always-on headless Service alongside RayMultiHostIndexing, or
2. Keep this approach: headless Service by default for all worker groups, hostname/subdomain still tied to RayMultiHostIndexing (default on).

Note: The multi_host e2e test has been extended to support FQDN verification , depending on above decision will split into its own e2e test.

Related issue number

Closes #5000

Checks

• I've made sure the tests are passing.
• Testing Strategy
  • Unit tests
  • Manual tests
  • This PR is not tested :(

Manual test instructions

1. Build/deploy operator with this change (./hack/local_deploy.sh or equivalent) and ensure RayMultiHostIndexing=true.
2. Apply a sample cluster:

   kubectl apply -f ray-operator/config/samples/ray-cluster.sample.yaml

3. Confirm worker hostname/subdomain and headless Service:

   kubectl get svc | grep headless
   kubectl get pods -l ray.io/node-type=worker -o custom-columns=\
   NAME:.metadata.name,HOSTNAME:.spec.hostname,SUBDOMAIN:.spec.subdomain,IP:.status.podIP

4. DNS lookup:

   kubectl run dnscheck --rm --restart=Never --image=busybox:1.36 --attach -- \
     nslookup workergroup-0.raycluster-kuberay-headless.default.svc.cluster.local

5. Delete the worker Pod and confirm the replacement keeps the same hostname/FQDN with a new IP/UID.
6. Unit tests:

   cd ray-operator
   go test ./controllers/ray/common/ -run TestDeafultWorkerPodTemplateWithReplicaGrpAndIndex -count=1
   go test ./controllers/ray/ -run TestReconcileHeadlessService -count=1

7. E2E (operator image must include this change):

   cd ray-operator
   go test -timeout 30m -v ./test/e2e -run TestRayClusterSingleHostMultiSlice
   go test -timeout 60m -v ./test/e2e -run TestRayClusterMultiHostMultiSlice

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{Reviewed by Cursor Bugbot for commit 0b6dd32. Configure here.}

[cursor]

Invalid DNS worker hostnames

High Severity

generateWorkerHostname builds pod.spec.hostname from raw groupName and only runs CheckLabel, which does not enforce DNS-1123 rules. Worker groupName values may include uppercase or _ today, and with RayMultiHostIndexing defaulting on those pods are rejected on create. Existing clusters that previously worked can stop reconciling workers.

Additional Locations (1)

• ray-operator/controllers/ray/common/pod.go#L690-L691

 

^{Reviewed by Cursor Bugbot for commit 0b6dd32. Configure here.}