Order nspawn after required host systemd services

[Patrick W. Healy (phealy)]

Summary

• order the AKS Flex Node agent and nspawn machine after systemd-udev-settle.service
• add bootstrap.additionalRequiredServices for host services that must complete before nspawn starts
• render configured services into the per-machine nspawn drop-in as both Requires= and After= dependencies
• validate configured values as systemd .service unit names and document the new setting

Motivation

Some host images run services such as ib_rdma_configure.service that rename or otherwise reconfigure devices during boot. If nspawn starts concurrently, it can capture stale device names and fail to expose the expected objects to the worker node. Explicit ordering ensures host device discovery and renaming finish before the machine starts.

Configuration

{
  "bootstrap": {
    "additionalRequiredServices": [
      "ib_rdma_configure.service"
    ]
  }
}

The generated nspawn drop-in requires each configured service and orders nspawn after it. The dependency configuration is reconciled during bootstrap and repave and remains safe to apply repeatedly.

Validation

• make check
• focused config and daemon tests, including invalid unit names and idempotent drop-in reconciliation

[Copilot]

Pull request overview

This PR enhances AKS Flex Node’s host-boot ordering guarantees by adding explicit systemd dependency wiring so the agent and each systemd-nspawn machine can be started after udev settle and optionally after/with additional host services that must finish device setup before the containerized node starts.

Changes:

• Add a daemon task to render a per-machine systemd-nspawn drop-in with After= and (optionally) Requires= dependencies, and reload systemd.
• Introduce bootstrap.additionalRequiredServices config with validation for safe .service unit names, plus tests.
• Update the agent’s systemd unit to order after systemd-udev-settle.service, plus docs and lifecycle assertions.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
pkg/daemon/systemd_dependencies.go	New task that writes the nspawn dependency drop-in and reloads systemd.
pkg/daemon/systemd_dependencies_test.go	Tests drop-in rendering behavior and repeated application.
pkg/daemon/start.go	Wires the new dependency task into the node start sequence.
pkg/daemon/lifecycle_test.go	Asserts the agent unit orders after systemd-udev-settle.service.
pkg/daemon/assets/aks-flex-node-agent.service	Updates agent unit ordering to include systemd-udev-settle.service.
pkg/config/config.go	Adds bootstrap.additionalRequiredServices and validates unit-name format.
pkg/config/config_test.go	Adds coverage for config load + validation of required services.
docs/usages/configuration.md	Documents the new configuration field and its purpose.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated no new comments.