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
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
|
/*
Copyright The containerd Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package cgroups
import (
"fmt"
"io/ioutil"
"os"
"path/filepath"
"strconv"
"strings"
"sync"
specs "github.com/opencontainers/runtime-spec/specs-go"
"github.com/pkg/errors"
)
// New returns a new control via the cgroup cgroups interface
func New(hierarchy Hierarchy, path Path, resources *specs.LinuxResources) (Cgroup, error) {
subsystems, err := hierarchy()
if err != nil {
return nil, err
}
for _, s := range subsystems {
if err := initializeSubsystem(s, path, resources); err != nil {
return nil, err
}
}
return &cgroup{
path: path,
subsystems: subsystems,
}, nil
}
// Load will load an existing cgroup and allow it to be controlled
func Load(hierarchy Hierarchy, path Path) (Cgroup, error) {
subsystems, err := hierarchy()
if err != nil {
return nil, err
}
// check the the subsystems still exist
for _, s := range pathers(subsystems) {
p, err := path(s.Name())
if err != nil {
if os.IsNotExist(errors.Cause(err)) {
return nil, ErrCgroupDeleted
}
return nil, err
}
if _, err := os.Lstat(s.Path(p)); err != nil {
if os.IsNotExist(err) {
return nil, ErrCgroupDeleted
}
return nil, err
}
}
return &cgroup{
path: path,
subsystems: subsystems,
}, nil
}
type cgroup struct {
path Path
subsystems []Subsystem
mu sync.Mutex
err error
}
// New returns a new sub cgroup
func (c *cgroup) New(name string, resources *specs.LinuxResources) (Cgroup, error) {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return nil, c.err
}
path := subPath(c.path, name)
for _, s := range c.subsystems {
if err := initializeSubsystem(s, path, resources); err != nil {
return nil, err
}
}
return &cgroup{
path: path,
subsystems: c.subsystems,
}, nil
}
// Subsystems returns all the subsystems that are currently being
// consumed by the group
func (c *cgroup) Subsystems() []Subsystem {
return c.subsystems
}
// Add moves the provided process into the new cgroup
func (c *cgroup) Add(process Process) error {
if process.Pid <= 0 {
return ErrInvalidPid
}
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
return c.add(process)
}
func (c *cgroup) add(process Process) error {
for _, s := range pathers(c.subsystems) {
p, err := c.path(s.Name())
if err != nil {
return err
}
if err := ioutil.WriteFile(
filepath.Join(s.Path(p), cgroupProcs),
[]byte(strconv.Itoa(process.Pid)),
defaultFilePerm,
); err != nil {
return err
}
}
return nil
}
// AddTask moves the provided tasks (threads) into the new cgroup
func (c *cgroup) AddTask(process Process) error {
if process.Pid <= 0 {
return ErrInvalidPid
}
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
return c.addTask(process)
}
func (c *cgroup) addTask(process Process) error {
for _, s := range pathers(c.subsystems) {
p, err := c.path(s.Name())
if err != nil {
return err
}
if err := ioutil.WriteFile(
filepath.Join(s.Path(p), cgroupTasks),
[]byte(strconv.Itoa(process.Pid)),
defaultFilePerm,
); err != nil {
return err
}
}
return nil
}
// Delete will remove the control group from each of the subsystems registered
func (c *cgroup) Delete() error {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
var errors []string
for _, s := range c.subsystems {
if d, ok := s.(deleter); ok {
sp, err := c.path(s.Name())
if err != nil {
return err
}
if err := d.Delete(sp); err != nil {
errors = append(errors, string(s.Name()))
}
continue
}
if p, ok := s.(pather); ok {
sp, err := c.path(s.Name())
if err != nil {
return err
}
path := p.Path(sp)
if err := remove(path); err != nil {
errors = append(errors, path)
}
}
}
if len(errors) > 0 {
return fmt.Errorf("cgroups: unable to remove paths %s", strings.Join(errors, ", "))
}
c.err = ErrCgroupDeleted
return nil
}
// Stat returns the current metrics for the cgroup
func (c *cgroup) Stat(handlers ...ErrorHandler) (*Metrics, error) {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return nil, c.err
}
if len(handlers) == 0 {
handlers = append(handlers, errPassthrough)
}
var (
stats = &Metrics{
CPU: &CPUStat{
Throttling: &Throttle{},
Usage: &CPUUsage{},
},
}
wg = &sync.WaitGroup{}
errs = make(chan error, len(c.subsystems))
)
for _, s := range c.subsystems {
if ss, ok := s.(stater); ok {
sp, err := c.path(s.Name())
if err != nil {
return nil, err
}
wg.Add(1)
go func() {
defer wg.Done()
if err := ss.Stat(sp, stats); err != nil {
for _, eh := range handlers {
if herr := eh(err); herr != nil {
errs <- herr
}
}
}
}()
}
}
wg.Wait()
close(errs)
for err := range errs {
return nil, err
}
return stats, nil
}
// Update updates the cgroup with the new resource values provided
//
// Be prepared to handle EBUSY when trying to update a cgroup with
// live processes and other operations like Stats being performed at the
// same time
func (c *cgroup) Update(resources *specs.LinuxResources) error {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
for _, s := range c.subsystems {
if u, ok := s.(updater); ok {
sp, err := c.path(s.Name())
if err != nil {
return err
}
if err := u.Update(sp, resources); err != nil {
return err
}
}
}
return nil
}
// Processes returns the processes running inside the cgroup along
// with the subsystem used, pid, and path
func (c *cgroup) Processes(subsystem Name, recursive bool) ([]Process, error) {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return nil, c.err
}
return c.processes(subsystem, recursive)
}
func (c *cgroup) processes(subsystem Name, recursive bool) ([]Process, error) {
s := c.getSubsystem(subsystem)
sp, err := c.path(subsystem)
if err != nil {
return nil, err
}
path := s.(pather).Path(sp)
var processes []Process
err = filepath.Walk(path, func(p string, info os.FileInfo, err error) error {
if err != nil {
return err
}
if !recursive && info.IsDir() {
if p == path {
return nil
}
return filepath.SkipDir
}
dir, name := filepath.Split(p)
if name != cgroupProcs {
return nil
}
procs, err := readPids(dir, subsystem)
if err != nil {
return err
}
processes = append(processes, procs...)
return nil
})
return processes, err
}
// Freeze freezes the entire cgroup and all the processes inside it
func (c *cgroup) Freeze() error {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
s := c.getSubsystem(Freezer)
if s == nil {
return ErrFreezerNotSupported
}
sp, err := c.path(Freezer)
if err != nil {
return err
}
return s.(*freezerController).Freeze(sp)
}
// Thaw thaws out the cgroup and all the processes inside it
func (c *cgroup) Thaw() error {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
s := c.getSubsystem(Freezer)
if s == nil {
return ErrFreezerNotSupported
}
sp, err := c.path(Freezer)
if err != nil {
return err
}
return s.(*freezerController).Thaw(sp)
}
// OOMEventFD returns the memory cgroup's out of memory event fd that triggers
// when processes inside the cgroup receive an oom event. Returns
// ErrMemoryNotSupported if memory cgroups is not supported.
func (c *cgroup) OOMEventFD() (uintptr, error) {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return 0, c.err
}
s := c.getSubsystem(Memory)
if s == nil {
return 0, ErrMemoryNotSupported
}
sp, err := c.path(Memory)
if err != nil {
return 0, err
}
return s.(*memoryController).OOMEventFD(sp)
}
// State returns the state of the cgroup and its processes
func (c *cgroup) State() State {
c.mu.Lock()
defer c.mu.Unlock()
c.checkExists()
if c.err != nil && c.err == ErrCgroupDeleted {
return Deleted
}
s := c.getSubsystem(Freezer)
if s == nil {
return Thawed
}
sp, err := c.path(Freezer)
if err != nil {
return Unknown
}
state, err := s.(*freezerController).state(sp)
if err != nil {
return Unknown
}
return state
}
// MoveTo does a recursive move subsystem by subsystem of all the processes
// inside the group
func (c *cgroup) MoveTo(destination Cgroup) error {
c.mu.Lock()
defer c.mu.Unlock()
if c.err != nil {
return c.err
}
for _, s := range c.subsystems {
processes, err := c.processes(s.Name(), true)
if err != nil {
return err
}
for _, p := range processes {
if err := destination.Add(p); err != nil {
return err
}
}
}
return nil
}
func (c *cgroup) getSubsystem(n Name) Subsystem {
for _, s := range c.subsystems {
if s.Name() == n {
return s
}
}
return nil
}
func (c *cgroup) checkExists() {
for _, s := range pathers(c.subsystems) {
p, err := c.path(s.Name())
if err != nil {
return
}
if _, err := os.Lstat(s.Path(p)); err != nil {
if os.IsNotExist(err) {
c.err = ErrCgroupDeleted
return
}
}
}
}
|