ananke/internal/cluster/orchestrator_inventory_test.go

package cluster

import (
	"context"
	"log"
	"os"
	"path/filepath"
	"strings"
	"testing"

	"scm.bstein.dev/bstein/ananke/internal/config"
	"scm.bstein.dev/bstein/ananke/internal/execx"
)

// TestValidateNodeInventoryPassesForStrictMappings runs one orchestration or CLI step.
// Signature: TestValidateNodeInventoryPassesForStrictMappings(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestValidateNodeInventoryPassesForStrictMappings(t *testing.T) {
	orch := &Orchestrator{
		cfg: config.Config{
			SSHUser: "atlas",
			SSHPort: 2277,
			SSHNodeHosts: map[string]string{
				"titan-0a": "192.168.22.11",
				"titan-0b": "192.168.22.12",
				"titan-0c": "192.168.22.13",
				"titan-22": "192.168.22.22",
			},
			SSHManagedNodes: []string{"titan-0a", "titan-0b", "titan-0c", "titan-22"},
			ControlPlanes:   []string{"titan-0a", "titan-0b", "titan-0c"},
			Workers:         []string{"titan-22"},
		},
		log: log.New(os.Stdout, "", 0),
	}
	if err := orch.validateNodeInventory(); err != nil {
		t.Fatalf("expected inventory to pass, got error: %v", err)
	}
}

// TestValidateNodeInventoryFailsWhenNodeMappingMissing runs one orchestration or CLI step.
// Signature: TestValidateNodeInventoryFailsWhenNodeMappingMissing(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestValidateNodeInventoryFailsWhenNodeMappingMissing(t *testing.T) {
	orch := &Orchestrator{
		cfg: config.Config{
			SSHUser:         "atlas",
			SSHPort:         2277,
			SSHNodeHosts:    map[string]string{"titan-0a": "192.168.22.11"},
			SSHManagedNodes: []string{"titan-0a", "titan-0b"},
			ControlPlanes:   []string{"titan-0a"},
			Workers:         []string{"titan-0b"},
		},
		log: log.New(os.Stdout, "", 0),
	}
	err := orch.validateNodeInventory()
	if err == nil {
		t.Fatalf("expected inventory error for missing mapping")
	}
	if !strings.Contains(err.Error(), "missing ssh_node_hosts entry") {
		t.Fatalf("expected missing-mapping detail, got: %v", err)
	}
}

// TestValidateNodeInventoryFailsWhenWorkerNotManaged runs one orchestration or CLI step.
// Signature: TestValidateNodeInventoryFailsWhenWorkerNotManaged(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestValidateNodeInventoryFailsWhenWorkerNotManaged(t *testing.T) {
	orch := &Orchestrator{
		cfg: config.Config{
			SSHUser: "atlas",
			SSHPort: 2277,
			SSHNodeHosts: map[string]string{
				"titan-0a": "192.168.22.11",
				"titan-22": "192.168.22.22",
			},
			SSHManagedNodes: []string{"titan-0a"},
			ControlPlanes:   []string{"titan-0a"},
			Workers:         []string{"titan-22"},
		},
		log: log.New(os.Stdout, "", 0),
	}
	err := orch.validateNodeInventory()
	if err == nil {
		t.Fatalf("expected inventory error for unmanaged worker")
	}
	if !strings.Contains(err.Error(), "missing from ssh_managed_nodes") {
		t.Fatalf("expected unmanaged-worker detail, got: %v", err)
	}
}

// TestNormalizeShutdownModeRejectsPoweroff runs one orchestration or CLI step.
// Signature: TestNormalizeShutdownModeRejectsPoweroff(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestNormalizeShutdownModeRejectsPoweroff(t *testing.T) {
	if _, err := normalizeShutdownMode("poweroff"); err == nil {
		t.Fatalf("expected poweroff mode to be rejected")
	}
	if mode, err := normalizeShutdownMode("config"); err != nil || mode != "cluster-only" {
		t.Fatalf("expected config mode to normalize to cluster-only, mode=%q err=%v", mode, err)
	}
}

// TestWaitForNodeInventoryReachabilityPasses runs one orchestration or CLI step.
// Signature: TestWaitForNodeInventoryReachabilityPasses(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestWaitForNodeInventoryReachabilityPasses(t *testing.T) {
	tmp := t.TempDir()
	sshPath := filepath.Join(tmp, "ssh")
	sshScript := `#!/usr/bin/env bash
set -euo pipefail
target=""
for arg in "$@"; do
  case "$arg" in
    *@*) target="$arg" ;;
  esac
done
if [[ "$target" == *"titan-0a"* ]] || [[ "$target" == *"titan-0b"* ]]; then
  echo "__ANANKE_NODE_REACHABLE__"
  exit 0
fi
echo "no route to host" >&2
exit 255
`
	if err := os.WriteFile(sshPath, []byte(sshScript), 0o755); err != nil {
		t.Fatalf("write fake ssh: %v", err)
	}
	t.Setenv("PATH", tmp+":"+os.Getenv("PATH"))

	orch := &Orchestrator{
		cfg: config.Config{
			SSHUser:         "atlas",
			SSHPort:         2277,
			ControlPlanes:   []string{"titan-0a"},
			Workers:         []string{"titan-0b"},
			SSHManagedNodes: []string{"titan-0a", "titan-0b"},
			Startup: config.Startup{
				RequireNodeInventoryReach:     true,
				NodeInventoryReachWaitSeconds: 2,
				NodeInventoryReachPollSeconds: 1,
				IgnoreUnavailableNodes:        []string{},
			},
		},
		runner: &execx.Runner{DryRun: false},
		log:    log.New(os.Stdout, "", 0),
	}
	if err := orch.waitForNodeInventoryReachability(context.Background()); err != nil {
		t.Fatalf("expected node inventory reachability to pass, got: %v", err)
	}
}

// TestWaitForNodeInventoryReachabilityFails runs one orchestration or CLI step.
// Signature: TestWaitForNodeInventoryReachabilityFails(t *testing.T).
// Why: keeps behavior explicit so startup/shutdown workflows remain maintainable as services evolve.
func TestWaitForNodeInventoryReachabilityFails(t *testing.T) {
	tmp := t.TempDir()
	sshPath := filepath.Join(tmp, "ssh")
	sshScript := `#!/usr/bin/env bash
set -euo pipefail
echo "no route to host" >&2
exit 255
`
	if err := os.WriteFile(sshPath, []byte(sshScript), 0o755); err != nil {
		t.Fatalf("write fake ssh: %v", err)
	}
	t.Setenv("PATH", tmp+":"+os.Getenv("PATH"))

	orch := &Orchestrator{
		cfg: config.Config{
			SSHUser:         "atlas",
			SSHPort:         2277,
			ControlPlanes:   []string{"titan-0a"},
			SSHManagedNodes: []string{"titan-0a"},
			Startup: config.Startup{
				RequireNodeInventoryReach:     true,
				NodeInventoryReachWaitSeconds: 1,
				NodeInventoryReachPollSeconds: 1,
			},
		},
		runner: &execx.Runner{DryRun: false},
		log:    log.New(os.Stdout, "", 0),
	}
	err := orch.waitForNodeInventoryReachability(context.Background())
	if err == nil {
		t.Fatalf("expected node inventory reachability to fail")
	}
	if !strings.Contains(err.Error(), "did not pass within") {
		t.Fatalf("expected timeout detail in error, got: %v", err)
	}
}