-
Notifications
You must be signed in to change notification settings - Fork 50
/
kernel.go
307 lines (242 loc) · 6.97 KB
/
kernel.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
package kernel
import (
"cmp"
"context"
"errors"
"fmt"
"os"
"os/signal"
"slices"
"sync"
"sync/atomic"
"time"
"github.com/justtrackio/gosoline/pkg/appctx"
"github.com/justtrackio/gosoline/pkg/cfg"
"github.com/justtrackio/gosoline/pkg/coffin"
"github.com/justtrackio/gosoline/pkg/conc"
"github.com/justtrackio/gosoline/pkg/log"
"golang.org/x/sys/unix"
)
const (
ExitCodeOk = 0
ExitCodeErr = 1
ExitCodeNothingToRun = 10
ExitCodeNoForeground = 11
ExitCodeForced = 12
)
type ExitHandler func(code int)
type Settings struct {
KillTimeout time.Duration `cfg:"kill_timeout" default:"10s"`
HealthCheck HealthCheckSettings `cfg:"health_check"`
}
//go:generate mockery --name Kernel
type Kernel interface {
HealthCheck() HealthCheckResult
Running() <-chan struct{}
Run()
Stop(reason string)
}
type kernelOption func(k *kernel)
type kernel struct {
ctx context.Context
config cfg.Config
logger log.Logger
middlewares []Middleware
stages stages
running chan struct{}
stopOnce sync.Once
foregroundModules int32
killTimeout time.Duration
exitCode int
exitOnce sync.Once
exitHandler ExitHandler
}
func newKernel(ctx context.Context, config cfg.Config, logger log.Logger) (*kernel, error) {
settings := &Settings{}
config.UnmarshalKey("kernel", settings)
k := &kernel{
config: config,
logger: logger.WithChannel("kernel"),
ctx: ctx,
running: make(chan struct{}),
killTimeout: settings.KillTimeout,
exitCode: ExitCodeErr,
exitHandler: os.Exit,
}
_, err := appctx.Provide(ctx, healthCheckerKey, func() (HealthChecker, error) {
return k.HealthCheck, nil
})
return k, err
}
func (k *kernel) init(middlewares []Middleware, stages map[int]*stage) {
k.middlewares = middlewares
k.stages = stages
}
// Run will boot and run the modules added to the kernel.
// By default, os.Exit will get called if an error occurs or after the modules have stopped running,
// which means that there will be no return out of this call.
func (k *kernel) Run() {
defer k.exit()
k.logger.Info("starting kernel")
k.foregroundModules = k.stages.countForegroundModules()
k.debugConfig()
sig := make(chan os.Signal, 2)
signal.Notify(sig, unix.SIGTERM, unix.SIGINT)
go func() {
signal := <-sig
reason := fmt.Sprintf("signal %s", signal.String())
k.Stop(reason)
}()
runHandler := func() {
if err := k.runStages(); err != nil {
reason := fmt.Sprintf("error during running all stages: %s", err)
k.Stop(reason)
}
k.logger.Info("kernel up and running")
close(k.running)
<-k.waitAllStagesDone().Channel()
k.Stop("context done")
k.waitStopped()
hasErr := false
for _, stage := range k.stages {
if stage.err != nil && !errors.Is(stage.err, ErrKernelStopping) {
hasErr = true
}
}
if !hasErr {
k.exitCode = ExitCodeOk
}
}
for i := len(k.middlewares) - 1; i >= 0; i-- {
runHandler = k.middlewares[i](runHandler)
}
runHandler()
}
func (k *kernel) Stop(reason string) {
k.stopOnce.Do(func() {
go func() {
k.logger.Info("stopping kernel due to: %s", reason)
indices := k.stages.getIndices()
for i := len(indices) - 1; i >= 0; i-- {
stageIndex := indices[i]
k.logger.Info("stopping stage %d", stageIndex)
k.stages[stageIndex].stopWait()
k.logger.Info("stopped stage %d", stageIndex)
}
}()
})
}
func (k *kernel) Running() <-chan struct{} {
return k.running
}
func (k *kernel) HealthCheck() HealthCheckResult {
var result HealthCheckResult
for _, stageIndex := range k.stages.getIndices() {
stageResult := k.stages[stageIndex].healthcheck()
result = append(result, stageResult...)
}
slices.SortFunc(result, func(a, b ModuleHealthCheckResult) int {
return cmp.Compare(a.StageIndex, b.StageIndex)
})
return result
}
func (k *kernel) exit() {
k.exitOnce.Do(func() {
k.logger.Info("leaving kernel with exit code %d", k.exitCode)
k.exitHandler(k.exitCode)
})
}
func (k *kernel) debugConfig() {
debugErr := cfg.DebugConfig(k.config, k.logger)
if debugErr != nil {
k.logger.Error("can not debug config: %w", debugErr)
}
}
func (k *kernel) runStages() error {
for _, stageIndex := range k.stages.getIndices() {
if err := k.stages[stageIndex].run(k); err != nil {
return fmt.Errorf("can not run stage %d: %w", stageIndex, err)
}
k.logger.Info("stage %d up and running with %d modules", stageIndex, k.stages[stageIndex].len())
}
return nil
}
func (k *kernel) runModule(ctx context.Context, name string, ms *moduleState) (moduleErr error) {
defer k.logger.Info("stopped %s module %s", ms.config.GetType(), name)
k.logger.Info("running %s module %s in stage %d", ms.config.GetType(), name, ms.config.stage)
atomic.StoreInt32(&ms.isRunning, 1)
defer func(ms *moduleState) {
// recover any crash from the module - if we let the coffin handle this,
// this is already too late because we might have killed the kernel and
// swallowed the error
panicErr := coffin.ResolveRecovery(recover())
if panicErr != nil {
ms.err = panicErr
}
if ms.err != nil {
k.logger.Error("error running %s module %s: %w", ms.config.GetType(), name, ms.err)
}
atomic.StoreInt32(&ms.isRunning, 0)
if ms.config.essential {
k.essentialModuleExited(name)
} else if !ms.config.background {
k.foregroundModuleExited()
}
// make sure we are returning the correct error to our caller
moduleErr = ms.err
}(ms)
ms.err = ms.module.Run(ctx)
return ms.err
}
func (k *kernel) essentialModuleExited(name string) {
// actually we would need to decrement k.foregroundModules here, too
// however, as we are stopping in any case, we don't have to
reason := fmt.Sprintf("the essential module [%s] has stopped running", name)
k.Stop(reason)
}
func (k *kernel) foregroundModuleExited() {
remaining := atomic.AddInt32(&k.foregroundModules, -1)
if remaining == 0 {
k.Stop("no more foreground modules in running state")
}
}
func (k *kernel) waitStopped() {
done := conc.NewSignalOnce()
defer done.Signal()
go func() {
timer := time.NewTimer(k.killTimeout)
defer timer.Stop()
select {
case <-timer.C:
err := fmt.Errorf("kernel was not able to shutdown in %v", k.killTimeout)
k.logger.Error("kernel shutdown seems to be blocking.. exiting...: %w", err)
// we don't need to iterate in order, but doing so is much nicer, so lets do it
for _, stageIndex := range k.stages.getIndices() {
s := k.stages[stageIndex]
for name, ms := range s.modules.modules {
if atomic.LoadInt32(&ms.isRunning) != 0 {
k.logger.Info("module in stage %d blocking the shutdown: %s", stageIndex, name)
}
}
}
k.exitCode = ExitCodeForced
k.exit()
case <-done.Channel():
return
}
}()
// we don't need to iterate in order, we just need to block until everything is done
for _, stage := range k.stages {
<-stage.terminated.Channel()
}
}
func (k *kernel) waitAllStagesDone() conc.SignalOnce {
done := conc.NewSignalOnce()
go func() {
for _, s := range k.stages {
<-s.ctx.Done()
}
done.Signal()
}()
return done
}