Files
pangolin/server/internal/provision/service_test.go
T
wangjia 3d5bac66b4 feat(provision): 弹性节点基建 Terraform + 一键更换 (tsk_6u0FxmbC7Yeq)
IaC 面 (infra/) 与控制面 (server/internal/provision) 双产出,落地 doc/04 §4
「节点是牲口」弹性拓扑与 make-before-break 一键更换。

server/internal/provision:
- CloudAdapter 适配层 + Registry(首发 vultr 消耗品池 / hetzner 精品池各一);
  厂商凭证仅从 PROVISION_<VENDOR>_* env 注入,不入库不入 git。
- ProvisionService:CreateNode(幂等键重放不重复开机)、DestroyNode(幂等)、
  RotateIP(换 IP 不换机 + version bump)、ListProviders。
- Replace 一键更换:先建后拆,新机 up 先于旧机 draining(容量不下降),
  replacement_uuid 幂等键 + replacements 表分步记录,崩溃可续跑不重复。
- RotatePool:池内滚动轮换,并发度 1–2。
- cmd/nodectl CLI:create/destroy/rotate-ip/replace/rotate-pool/providers。
- 单测(mock 厂商 API + 内存 Store):幂等重放、make-before-break 时序断言、
  开机失败/探活超时→destroyed+失败计数+告警钩子、崩溃续跑、RotatePool。

infra/:
- terraform/:探针机 + 控制面基线模块化(probe / control-plane)+ README,
  低频基线进 state,节点不进 Terraform。
- cloud-init/node.yaml.tmpl:节点引导模板(注入一次性 bootstrap token,task #5)。
- identity-isolation.md:身份隔离登记表(doc/06 §2 红线,无任何凭证)。

migrations/000008:nodes 增 provider_instance_id/elastic_ip_id、node_events
增 ip_rotated、provision_idempotency / replacements 表(附加式,不动现网)。

红线:仅面向新厂商池,绝不纳管现网生产 EC2(deploy/ marzban)。

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-13 14:23:39 +08:00

156 lines
4.3 KiB
Go

package provision
import (
"context"
"errors"
"testing"
)
func TestCreateNode_Success(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
n, err := h.svc.CreateNode(ctx, freeSpec(), "key-1")
if err != nil {
t.Fatalf("CreateNode: %v", err)
}
if n.Status != StatusProvisioning {
t.Errorf("status = %s, want provisioning", n.Status)
}
if n.ProviderInstanceID == "" {
t.Error("instance ID not recorded")
}
if n.Endpoint == pendingEndpoint || n.Endpoint == "" {
t.Errorf("endpoint not updated: %q", n.Endpoint)
}
if h.tokens.calls != 1 {
t.Errorf("bootstrap token issued %d times, want 1", h.tokens.calls)
}
if h.adapter.createCount() != 1 {
t.Errorf("boot calls = %d, want 1", h.adapter.createCount())
}
if h.store.countEvents(EventProvisioned) != 1 {
t.Errorf("provisioned events = %d, want 1", h.store.countEvents(EventProvisioned))
}
}
func TestCreateNode_IdempotentReplay(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
n1, err := h.svc.CreateNode(ctx, freeSpec(), "same-key")
if err != nil {
t.Fatalf("first CreateNode: %v", err)
}
n2, err := h.svc.CreateNode(ctx, freeSpec(), "same-key")
if err != nil {
t.Fatalf("replay CreateNode: %v", err)
}
if n1.UUID != n2.UUID {
t.Errorf("replay returned different node: %s vs %s", n1.UUID, n2.UUID)
}
if h.adapter.createCount() != 1 {
t.Errorf("idempotency violated: boot called %d times, want 1", h.adapter.createCount())
}
}
func TestCreateNode_BootFailure(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
h.adapter.createErr = errors.New("vendor 500")
_, err := h.svc.CreateNode(ctx, freeSpec(), "key-boom")
if err == nil {
t.Fatal("expected error on boot failure")
}
// Node should be destroyed.
n, _ := h.store.GetNodeByUUID(ctx, h.store.idem["key-boom"])
if n == nil || n.Status != StatusDestroyed {
t.Errorf("node not destroyed after boot failure: %+v", n)
}
if got := h.svc.FailCount(PoolConsumable); got != 1 {
t.Errorf("fail count = %d, want 1", got)
}
if h.alert.count(AlertBootFailed) != 1 {
t.Errorf("boot_failed alerts = %d, want 1", h.alert.count(AlertBootFailed))
}
}
func TestDestroyNode_Idempotent(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
n, err := h.svc.CreateNode(ctx, freeSpec(), "key-d")
if err != nil {
t.Fatalf("CreateNode: %v", err)
}
if err := h.svc.DestroyNode(ctx, n.ID); err != nil {
t.Fatalf("DestroyNode: %v", err)
}
if err := h.svc.DestroyNode(ctx, n.ID); err != nil {
t.Fatalf("second DestroyNode should be no-op: %v", err)
}
if h.adapter.destroyCount() != 1 {
t.Errorf("vendor destroy called %d times, want 1 (idempotent)", h.adapter.destroyCount())
}
got, _ := h.store.GetNode(ctx, n.ID)
if got.Status != StatusDestroyed {
t.Errorf("status = %s, want destroyed", got.Status)
}
}
func TestRotateIP(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
n, err := h.svc.CreateNode(ctx, freeSpec(), "key-ip")
if err != nil {
t.Fatalf("CreateNode: %v", err)
}
before := h.store.version
oldEndpoint := n.Endpoint
updated, err := h.svc.RotateIP(ctx, n.ID)
if err != nil {
t.Fatalf("RotateIP: %v", err)
}
if updated.Endpoint == oldEndpoint {
t.Errorf("endpoint not changed: still %s", updated.Endpoint)
}
if h.store.version <= before {
t.Errorf("directory version not bumped: %d <= %d", h.store.version, before)
}
if h.store.countEvents(EventIPRotated) != 1 {
t.Errorf("ip_rotated events = %d, want 1", h.store.countEvents(EventIPRotated))
}
stored, _ := h.store.GetNode(ctx, n.ID)
if stored.Endpoint != updated.Endpoint {
t.Errorf("stored endpoint %s != returned %s", stored.Endpoint, updated.Endpoint)
}
}
func TestRotateIP_Unsupported(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
h.adapter.elastic = false
n, err := h.svc.CreateNode(ctx, freeSpec(), "key-ip2")
if err != nil {
t.Fatalf("CreateNode: %v", err)
}
_, err = h.svc.RotateIP(ctx, n.ID)
if !errors.Is(err, ErrElasticIPUnsupported) {
t.Errorf("err = %v, want ErrElasticIPUnsupported", err)
}
}
func TestListProviders(t *testing.T) {
h := newHarness(t)
ctx := context.Background()
ps, err := h.svc.ListProviders(ctx, PoolPremium)
if err != nil {
t.Fatalf("ListProviders: %v", err)
}
if len(ps) != 1 || ps[0].Pool != PoolPremium {
t.Errorf("premium providers = %+v, want exactly 1 premium", ps)
}
}