fix(spur-k8s): give SPUR_PEER_NODES resolvable pod names

[pre]

Related:

• fix(spurctld)!: refuse to bootstrap a second raft cluster #806
• fix(spur-k8s): keep the cancel of a deleted SpurJob #808
• fix(k8s): recover a node the controller has forgotten #809
• fix(spurctld): stop serving when the raft core is dead #810

Motivation

In Pod mode SPUR_PEER_NODES is unusable, so no distributed workload can start.

Every Kubernetes node registers the single operator address as its agent address.
The peer list the controller dispatches therefore holds that one address repeated
once per node, and the variable reads
"<operator-ip>:6818,<operator-ip>:6818,<operator-ip>:6818". An MPI or torchrun
launcher that reads it addresses the operator three times and never reaches a
peer. A single node job shows the same value, so it was never usable in Pod mode.

Technical Details

The Pods of a multi-node job already resolve each other. launch_job sets the Pod
hostname to the target node name and the subdomain to the job's headless Service,
so <node>.<job-service>.<namespace>.svc.cluster.local is a working name for each
peer.

SPUR_PEER_NODES now carries those names, in nodelist order, so index N is the
peer whose SPUR_NODE_RANK is N. A small headless_peer_dns helper builds them.

A single node job has no headless Service and therefore no name to publish. The
variable is left unset rather than holding an address that resolves to nothing,
so a launcher fails loudly instead of hanging on a dead address.

This also corrects the message the operator returns when no SpurJob carries the
requested job id. It said "agent unreachable", although the operator is running and
answering. It now says that Pod mode launches only a SpurJob, and that a job from
sbatch or spur submit has nothing for the operator to launch.

Related:

• fix(spurctld)!: refuse to bootstrap a second raft cluster #806
• fix(spur-k8s): keep the cancel of a deleted SpurJob #808
• fix(k8s): recover a node the controller has forgotten #809
• fix(spurctld): stop serving when the raft core is dead #810

Test Plan

1. Run a three node job in Pod mode and read SPUR_PEER_NODES inside every Pod.
2. Resolve each name from inside a Pod and compare against the real Pod IPs.
3. Run a single node job and confirm the variable is absent.
4. Unit tests for the name builder, including the nodelist ordering.

Test Result

Environment: three node RKE2 cluster on cloud VMs, 8 vCPU and 96 GiB each, no GPU.

• All three Pods of a three node job read the same list of three names, in nodelist
  order. Each name resolved from inside a Pod to the Pod IP of the matching node.
• Before the change the same job read one address three times.
• A single node job left SPUR_PEER_NODES unset.
• cargo clippy --workspace --exclude spur-ffi --all-targets --locked reports
  nothing; cargo test --locked passes 3528 tests, including three new ones for
  the name builder.

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.