Files
HomeAgent/internal/plugins/pluginmgr/upgrade_test.go
JianFeeeee 02cc74ce11 fix(proc): 子进程崩溃自愈 + 集中台账 + 注册面摘除
根因:子进程插件被 kill 后,内核只发了一个无人订阅的事件,
工具/stage handler/IO 通道全留在注册表里指向死进程,
模型继续调用只吃 ErrProcessExited,没有任何路径把插件拉回来。

## 四层修复

### 1. 专职 waitLoop(进程收割)
- 每个子进程配一根 waitLoop goroutine,是 cmd.Wait() 的唯一调用点
- 不再依赖 stdout EOF 判定死亡(孙子进程继承 stdout 时 EOF 永不到来)
- 手工 os.Pipe 替代 cmd.StdinPipe/StdoutPipe,避免 waitLoop 与
  os/exec 的内部关闭竞争
- host.go: Host.Supervisor(),Host.Close() 先 StopAll 再拆段

### 2. 集中台账 Supervisor
- proc/supervisor.go: 插件 Spawn 握手成功即 track,进程退出即 untrack
- StopAll: 并发发 plugin.stop 走优雅路径,到期仍在的一律 Kill
- 关停后才完成握手的进程被立即结束,不会活过内核
- 消除「孤儿进程持共享段映射 → SIGBUS」的隐患

### 3. 注册面摘除(detachPlugin)
- 新增 StageHost.UnregisterPluginStages:摘除指定插件的全部 stage handler
- 新增 Registry.pluginChannels 台账:记录每个插件注册的 IO 通道
- 三条路径统一走 detachPlugin:Disable / ReloadOne / RemovePlugin
- StopAndUnload 漏了 IO 通道也一并补上

### 4. 自动重启
- onProcCrash 从「只发事件」改为「摘注册面 → 从注册表移除 → 异步排重启」
- scheduleProcRestart: 窗口 5 分钟内最多 3 次,线性退避 1s/2s/3s
- 超限停手留日志;重启前复核是否已被 Disable 或被其他路径加载
- 崩溃计数窗口过期自动归零

### 5. 主动停止 vs 崩溃的区分
- proc.Plugin 新增 stopping 标志:Stop()/Close() 里 Set(true)
- handleExit 读 stopping 标志,主动停止不上报 onCrash
- 防止重载/禁用/卸载被误判为崩溃触发多余重启

### 6. Linux Pdeathsig 兜底
- procattr_linux.go: SysProcAttr.Pdeathsig = SIGKILL
- 兜 homed 自身被 SIGKILL/OOM 时子进程变孤儿的场景
- macOS/Windows 无等价物,空实现

### 7. pluginmgr 升级
- PluginManager 接口新增 PluginRuntime / ListPluginRuntimes
- plugin_list 输出运行态:loaded / alive / pid / crash_count / channel
- 新增 plugin_status: 全量运行期快照 + dead/unhealthy 汇总
- 新增 plugin_restart: 无条件重启单个插件(plgreload 不动未改二进制的插件)

### 测试
- process_test.go: 3 例(grandchild stdout 感知 / Supervisor track-untrack /
  StopAll 无孤儿)
- crash_recovery_test.go: 8 例(detach 三项齐全 / 通道重注册 / 崩溃不阻塞 /
  退避阈值 / 窗口过期 / 关停中跳过 / PluginRuntime 通道识别)
- stages_plugin_test.go: 4 例(stage 按插件摘除 / 空 stage 清理 / 空名 no-op /
  工具+stage 双摘后可重新注册同名)
2026-09-03 12:37:58 +08:00

225 lines
7.9 KiB
Go
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

package pluginmgr
import (
"archive/zip"
"bytes"
"encoding/json"
"os"
"path/filepath"
"sync"
"testing"
"gitcode.com/JianFeeeee/HomeAgent/internal/events"
"gitcode.com/JianFeeeee/HomeAgent/internal/plugin"
sdk "gitcode.com/JianFeeeee/HomeAgent/internal/sdk"
)
func TestCmpVersion(t *testing.T) {
cases := []struct{ a, b string; want int }{
{"1.0.0", "1.0.0", 0},
{"1.0.1", "1.0.0", 1},
{"1.0.0", "1.0.1", -1},
{"1.0", "1.0.0", 0},
{"v2.0.0", "1.9.9", 1},
{"2.0.0", "10.0.0", -1}, // 数字比较而非字典序
{"1.0.0-alpha", "1.0.0", 0}, // 非数字段按 0
}
for _, c := range cases {
if got := cmpVersion(c.a, c.b); got != c.want {
t.Errorf("cmpVersion(%q,%q)=%d want %d", c.a, c.b, got, c.want)
}
}
}
// ---- 最小 mock SDK ----
type pmSettings struct{}
func (m *pmSettings) DataDir() string { return "/tmp/mock_data" }
func (m *pmSettings) Get(string) (interface{}, error) { return nil, nil }
func (m *pmSettings) Set(string, interface{}) error { return nil }
func (m *pmSettings) List(string) ([]string, error) { return nil, nil }
func (m *pmSettings) GetCore(string) (interface{}, error) { return nil, nil }
func (m *pmSettings) SetCore(string, interface{}) error { return nil }
func (m *pmSettings) ListCore(string) ([]string, error) { return nil, nil }
func (m *pmSettings) GetPlugin(string, string) (interface{}, error) { return nil, nil }
func (m *pmSettings) SetPlugin(string, string, interface{}) error { return nil }
func (m *pmSettings) ListPlugin(string, string) ([]string, error) { return nil, nil }
func (m *pmSettings) RegisterDef(sdk.ConfigDef) {}
func (m *pmSettings) Defs(string) []*sdk.ConfigDef { return nil }
func (m *pmSettings) Dump() map[string]interface{} { return nil }
func (m *pmSettings) Plugins() []string { return nil }
func (m *pmSettings) DefsCore(string) []*sdk.ConfigDef { return nil }
func (m *pmSettings) DefsPlugin(string, string) []*sdk.ConfigDef { return nil }
func (m *pmSettings) Remove(string) error { return nil }
func (m *pmSettings) RemoveCore(string) error { return nil }
func (m *pmSettings) RemovePlugin(string, string) error { return nil }
// fakePluginMgr 记录调用StopAndUnload 只记标志,不真正操作。
type fakePluginMgr struct {
mu sync.Mutex
stopAndUnloads []string
}
func (f *fakePluginMgr) ListLoadedPlugins() []string { return nil }
func (f *fakePluginMgr) ListDisabledPlugins() []sdk.DisabledPluginInfo { return nil }
func (f *fakePluginMgr) IsPluginDisabled(string) bool { return false }
func (f *fakePluginMgr) IsBuiltinPlugin(string) bool { return false }
func (f *fakePluginMgr) DisablePlugin(string, string) error { return nil }
func (f *fakePluginMgr) EnablePlugin(string) error { return nil }
func (f *fakePluginMgr) RemovePlugin(string) error { return nil }
func (f *fakePluginMgr) StopAndUnload(name string) error {
f.mu.Lock()
defer f.mu.Unlock()
f.stopAndUnloads = append(f.stopAndUnloads, name)
return nil
}
func (f *fakePluginMgr) ReloadPlugins() (string, error) { return "", nil }
func (f *fakePluginMgr) ReloadOne(string) error { return nil }
func (f *fakePluginMgr) PluginMetas() map[string]sdk.PluginMeta {
return map[string]sdk.PluginMeta{}
}
func (f *fakePluginMgr) PluginDir() string { return "" }
func (f *fakePluginMgr) PluginRuntime(string) (sdk.PluginRuntimeInfo, bool) {
return sdk.PluginRuntimeInfo{}, false
}
func (f *fakePluginMgr) ListPluginRuntimes() []sdk.PluginRuntimeInfo { return nil }
// buildHmap 构造一个最小 .hmap 包。
func buildHmap(t *testing.T, name, version string) []byte {
t.Helper()
var buf bytes.Buffer
zw := zip.NewWriter(&buf)
manifest := map[string]interface{}{
"name": name, "name_zh": name, "name_en": name,
"version": version, "entry": "plugin.bin",
}
mData, _ := json.Marshal(manifest)
f, _ := zw.Create("plugin.json")
f.Write(mData)
bin, _ := zw.Create("plugin.bin")
bin.Write([]byte("binary-" + name + "-" + version))
zw.Close()
return buf.Bytes()
}
func TestInstallThenUpgradeKeepsConfig(t *testing.T) {
dir := t.TempDir()
fm := &fakePluginMgr{}
bus := events.NewBus()
sdkInst := sdk.New("pluginmgr", sdk.SDKConfig{
Settings: &pmSettings{},
EventBus: bus,
PluginMgr: fm,
})
p := &Plugin{pluginDir: dir, sdk: sdkInst}
// 1. 首次安装 v1.0.0
r1, _ := p.installFromData(buildHmap(t, "demo", "1.0.0"), false)
m1 := r1.(map[string]interface{})
if m1["status"] != "installed" {
t.Fatalf("install failed: %v", m1)
}
if _, err := os.Stat(filepath.Join(dir, "demo", "plugin.json")); err != nil {
t.Fatalf("installed dir missing: %v", err)
}
// 2. 不带 overwrite 重装 → 报 already exists + remove_first hint
r2, _ := p.installFromData(buildHmap(t, "demo", "1.0.0"), false)
m2 := r2.(map[string]interface{})
if m2["error"] != "plugin already exists" || m2["hint"] == "" {
t.Fatalf("expected already-exists with hint, got %v", m2)
}
if m2["current"] != "1.0.0" {
t.Fatalf("current version not reported: %v", m2)
}
// 3. overwrite 升级 v1.0.0 → v2.0.0
r3, _ := p.installFromData(buildHmap(t, "demo", "2.0.0"), true)
m3 := r3.(map[string]interface{})
if m3["status"] != "installed" || m3["action"] != "upgraded" {
t.Fatalf("upgrade failed: %v", m3)
}
if m3["previous_version"] != "1.0.0" {
t.Fatalf("previous_version = %v", m3["previous_version"])
}
if m3["config_kept"] != true {
t.Fatalf("config_kept should be true: %v", m3)
}
// StopAndUnload 应被调用且不触发 RemovePlugin不删配置
fm.mu.Lock()
calls := append([]string{}, fm.stopAndUnloads...)
fm.mu.Unlock()
if len(calls) != 1 || calls[0] != "demo" {
t.Fatalf("StopAndUnload not called once with demo: %v", calls)
}
// 新二进制写入
binData, err := os.ReadFile(filepath.Join(dir, "demo", "plugin.bin"))
if err != nil {
t.Fatalf("read new bin: %v", err)
}
if string(binData) != "binary-demo-2.0.0" {
t.Fatalf("bin not overwritten: %q", string(binData))
}
// 4. 降级 v2.0.0 → v1.5.0
r4, _ := p.installFromData(buildHmap(t, "demo", "1.5.0"), true)
m4 := r4.(map[string]interface{})
if m4["action"] != "downgraded" {
t.Fatalf("downgrade action = %v", m4)
}
}
func TestExtractFailureRollsBack(t *testing.T) {
dir := t.TempDir()
fm := &fakePluginMgr{}
bus := events.NewBus()
sdkInst := sdk.New("pluginmgr", sdk.SDKConfig{
Settings: &pmSettings{},
EventBus: bus,
PluginMgr: fm,
})
p := &Plugin{pluginDir: dir, sdk: sdkInst}
// 先装 v1.0.0
if r, _ := p.installFromData(buildHmap(t, "rollback", "1.0.0"), false); r.(map[string]interface{})["status"] != "installed" {
t.Fatal("install failed")
}
// 构造损坏包zip 但缺 plugin.jsonextractPackage 会失败)
var buf bytes.Buffer
zw := zip.NewWriter(&buf)
f, _ := zw.Create("plugin.bin")
f.Write([]byte("corrupt"))
zw.Close()
// 畸形包在 validatePackage 层就拒绝,未达 extract——模拟 extract 失败:
// 直接注入非法平台文件触发 extractPackage 错误
bad := buildHmap(t, "rollback", "9.9.9")
// 篡改使 extract 失败:附加一个越界路径
var rb bytes.Buffer
zw2 := zip.NewWriter(&rb)
f2, _ := zw2.Create("../../evil")
f2.Write([]byte("x"))
mf, _ := zw2.Create("plugin.json")
mData, _ := json.Marshal(map[string]interface{}{"name": "rollback", "version": "9.9.9", "entry": "plugin.bin"})
mf.Write(mData)
zw2.Close()
bad = rb.Bytes()
r, _ := p.installFromData(bad, true)
m := r.(map[string]interface{})
if m["error"] == nil {
t.Fatalf("expected error for corrupt package, got %v", m)
}
if m["rollback"] != nil {
t.Fatalf("rollback itself failed: %v", m)
}
// 旧版应被恢复
mfest, err := plugin.ReadManifest(filepath.Join(dir, "rollback"))
if err != nil || mfest.Version != "1.0.0" {
t.Fatalf("old version not restored: %v / %v", mfest, err)
}
}