feat: run Docker launcher jobs across multiple nodes

[cquil11]

Stacked on #47 (single-node Docker launcher). This PR adds only the multi-node part: DockerLauncher runs a job's processes across several Docker hosts.

Design

• Nodes. docker_hosts in srtslurm.yaml lists the job's nodes and defaults to this machine. The orchestrator runs where srtctl apply runs and carves docker_hosts into roles with the same Nodes.from_nodelist a Slurm nodelist uses, so aggregated, multi-node aggregated and disaggregated recipes need no changes. validate() now checks the recipe's node count against docker_hosts; the single-node checks on dedicated frontend, benchmark or infra nodes and on service pools are removed, since with several hosts those are just more nodes.
• ssh fan-out. A LaunchSpec runs one task on each node in its nodelist. Each task is a docker run --gpus all --network host --ipc host container, started directly on this machine and through ssh -n -o BatchMode=yes <node> on any other node. A launch that targets several nodes (the exporters) starts one task per node in parallel, and %N in the log path becomes the node name, as with srun.
• Per-node IP lookup. With one host, every address is 127.0.0.1. With several, node_ip runs the existing ip_utils/get_node_ip.sh lookup on the target node over ssh, so network_interface behaves as it does under Slurm.
• Per-host kill and cleanup. A named step is stopped with docker kill --signal=... on its node, because killing the local ssh or docker run client does not stop the container. When the job script exits, any container labelled srtctl.job=<job_id> that is still running is removed on every host.
• Remote host processes. A container-less step on another node (a host-native exporter) runs under ssh -tt, so it gets SIGHUP when its ssh client is signalled; host steps are stopped by signalling their local process group. Without this, the remote process kept running after the job.
• Requirements. This machine needs passwordless ssh to every other host. Every path the job mounts (the model, outputs/, the srtctl checkout) must exist at the same path on every host, as on a Slurm cluster's shared filesystem.

Validation

Run on h200, on two nodes held with salloc (worker-12 and worker-13). Each node had Docker 29.2.1 and NVIDIA H200 GPUs. srtctl apply ran on worker-12 with launcher: docker. Model: Qwen3-0.6B (bf16). Image: lmsysorg/sglang:v0.5.20. Benchmark: sa-bench at ISL/OSL 128/128, concurrencies 4 and 8. All three runs used tree 51a61e2b, which is identical to this branch's tree after the merge of #47.

Cluster	Scenario	docker_hosts	Outcome	Notes
h200	Single-node agg (TP1)	worker-12	Pass	40/40 and 80/80 requests completed; no containers left
h200	2-node agg (one TP2 engine spanning both nodes)	worker-12, worker-13	Pass	--dist-init-addr set to worker-12's node IP; 40/40 and 80/80 requests completed; no containers or host exporters left on either node
h200	1P1D disagg (prefill on worker-12, decode on worker-13, sglang-router)	worker-12, worker-13	Pass	Mooncake KV transfer between the nodes; 40/40 and 80/80 requests completed; no containers left

A failed run (the prefill worker exited on a startup flag) stopped the job and removed both nodes' containers.

[github-actions]

src/srtctl diff: +151 -85