-
Notifications
You must be signed in to change notification settings - Fork 65
/
actuator_reconcile.go
442 lines (356 loc) · 14.1 KB
/
actuator_reconcile.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
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
// SPDX-FileCopyrightText: 2024 SAP SE or an SAP affiliate company and Gardener contributors
//
// SPDX-License-Identifier: Apache-2.0
package bastion
import (
"context"
"errors"
"fmt"
"time"
"github.com/aliyun/alibaba-cloud-sdk-go/services/ecs"
"github.com/gardener/gardener/extensions/pkg/controller"
"github.com/gardener/gardener/extensions/pkg/util"
extensionsv1alpha1 "github.com/gardener/gardener/pkg/apis/extensions/v1alpha1"
ctrlerror "github.com/gardener/gardener/pkg/controllerutils/reconciler"
"github.com/go-logr/logr"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/equality"
"sigs.k8s.io/controller-runtime/pkg/client"
"github.com/gardener/gardener-extension-provider-alicloud/pkg/alicloud"
aliclient "github.com/gardener/gardener-extension-provider-alicloud/pkg/alicloud/client"
"github.com/gardener/gardener-extension-provider-alicloud/pkg/apis/alicloud/helper"
)
// bastionEndpoints holds the endpoints the bastion host provides
type bastionEndpoints struct {
// private is the private endpoint of the bastion. It is required when opening a port on the worker node to allow SSH access from the bastion
private *corev1.LoadBalancerIngress
// public is the public endpoint where the end user connects to establish the SSH connection.
public *corev1.LoadBalancerIngress
}
// Ready returns true if both public and private interfaces each have either
// an IP or a hostname or both.
func (be *bastionEndpoints) Ready() bool {
return be != nil && IngressReady(be.private) && IngressReady(be.public)
}
func (a *actuator) Reconcile(ctx context.Context, log logr.Logger, bastion *extensionsv1alpha1.Bastion, cluster *controller.Cluster) error {
log.Info("Bastion reconciles operation")
opt, err := DetermineOptions(bastion, cluster)
if err != nil {
return err
}
infrastructureStatus, err := getInfrastructureStatus(ctx, a.client, cluster)
if err != nil {
return err
}
credentials, err := alicloud.ReadCredentialsFromSecretRef(ctx, a.client, &opt.SecretReference)
if err != nil {
return err
}
aliCloudECSClient, err := a.newClientFactory.NewECSClient(opt.Region, credentials.AccessKeyID, credentials.AccessKeySecret)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
imageID := infrastructureStatus.MachineImages[0].ID
vSwitchesZoneID := infrastructureStatus.VPC.VSwitches[0].Zone
vSwitchesID := infrastructureStatus.VPC.VSwitches[0].ID
vpcId := infrastructureStatus.VPC.ID
shootSecurityGroupId := infrastructureStatus.VPC.SecurityGroups[0].ID
var instanceTypeId string
for cores := 1; cores <= 2; cores++ {
instanceType, err := aliCloudECSClient.GetInstanceType(cores, vSwitchesZoneID)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
if instanceType == nil ||
len(instanceType.AvailableZones.AvailableZone) == 0 ||
len(instanceType.AvailableZones.AvailableZone[0].AvailableResources.AvailableResource) == 0 ||
len(instanceType.AvailableZones.AvailableZone[0].AvailableResources.AvailableResource[0].SupportedResources.SupportedResource) == 0 ||
instanceType.AvailableZones.AvailableZone[0].AvailableResources.AvailableResource[0].SupportedResources.SupportedResource[0].Status != "Available" {
continue
}
instanceTypeId = instanceType.AvailableZones.AvailableZone[0].AvailableResources.AvailableResource[0].SupportedResources.SupportedResource[0].Value
break
}
if instanceTypeId == "" {
if len(cluster.CloudProfile.Spec.MachineTypes) == 0 {
return errors.New("failed to determine instanceTypeId from cloud profile as fallback. Machine types missing from cloud profile")
}
instanceTypeId = cluster.CloudProfile.Spec.MachineTypes[0].Name
log.Info("falling back to first machine type of cloud profile as bastion instance type id", "instance type", cluster.CloudProfile.Spec.MachineTypes[0].Name)
}
securityGroupID, err := ensureSecurityGroup(aliCloudECSClient, opt.SecurityGroupName, vpcId, log)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
instanceID, err := ensureComputeInstance(aliCloudECSClient, log, opt, securityGroupID, imageID, vSwitchesID, vSwitchesZoneID, instanceTypeId)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
ready, err := isInstanceReady(aliCloudECSClient, opt)
if err != nil {
return util.DetermineError(fmt.Errorf("failed to check for bastion instance: %w", err), helper.KnownCodes)
}
if !ready {
return &ctrlerror.RequeueAfterError{
RequeueAfter: 10 * time.Second,
Cause: errors.New("bastion instance not ready yet"),
}
}
err = ensureSecurityGroupRules(aliCloudECSClient, opt, shootSecurityGroupId, bastion, securityGroupID)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
publicIP, err := aliCloudECSClient.AllocatePublicIp(instanceID)
if err != nil {
return util.DetermineError(err, helper.KnownCodes)
}
endpoints, err := getInstanceEndpoints(aliCloudECSClient, opt, publicIP.IpAddress)
if err != nil {
return err
}
if !endpoints.Ready() {
return &ctrlerror.RequeueAfterError{
// requeue rather soon, so that the user (most likely gardenctl eventually)
// doesn't have to wait too long for the public endpoint to become available
RequeueAfter: 5 * time.Second,
Cause: fmt.Errorf("bastion instance has no public/private endpoints yet"),
}
}
// once a public endpoint is available, publish the endpoint on the
// Bastion resource to notify upstream about the ready instance
patch := client.MergeFrom(bastion.DeepCopy())
bastion.Status.Ingress = endpoints.public
return a.client.Status().Patch(ctx, bastion, patch)
}
// IngressReady returns true if either an IP or a hostname or both are set.
func IngressReady(ingress *corev1.LoadBalancerIngress) bool {
return ingress != nil && (ingress.Hostname != "" || ingress.IP != "")
}
// addressToIngress converts the IP address into a
// corev1.LoadBalancerIngress resource. If both arguments are nil, then
// nil is returned.
func addressToIngress(dnsName *string, ipAddress *string) *corev1.LoadBalancerIngress {
var ingress *corev1.LoadBalancerIngress
if ipAddress != nil || dnsName != nil {
ingress = &corev1.LoadBalancerIngress{}
if dnsName != nil {
ingress.Hostname = *dnsName
}
if ipAddress != nil {
ingress.IP = *ipAddress
}
}
return ingress
}
func getInstanceEndpoints(c aliclient.ECS, opt *Options, ip string) (*bastionEndpoints, error) {
response, err := c.GetInstances(opt.BastionInstanceName)
if err != nil {
return nil, err
}
if response == nil {
return nil, fmt.Errorf("compute instance can't be nil")
}
if len(response.Instances.Instance) == 0 || response.Instances.Instance[0].Status != "Running" {
return nil, fmt.Errorf("compute instance not ready yet")
}
endpoints := &bastionEndpoints{}
instance := response.Instances.Instance[0]
internalIP := instance.VpcAttributes.PrivateIpAddress.IpAddress[0]
if ingress := addressToIngress(nil, &internalIP); ingress != nil {
endpoints.private = ingress
}
if ingress := addressToIngress(nil, &ip); ingress != nil {
endpoints.public = ingress
}
return endpoints, nil
}
func ensureComputeInstance(c aliclient.ECS, log logr.Logger, opt *Options, securityGroupID, imageID, vSwitchId, zoneID, instanceTypeID string) (string, error) {
response, err := c.GetInstances(opt.BastionInstanceName)
if err != nil {
return "", err
}
if len(response.Instances.Instance) > 0 && response.Instances.Instance[0].InstanceName == opt.BastionInstanceName {
return response.Instances.Instance[0].InstanceId, nil
}
log.Info("creating new bastion compute instance")
instance, err := c.CreateInstances(opt.BastionInstanceName, securityGroupID, imageID, vSwitchId, zoneID, instanceTypeID, opt.UserData)
if err != nil {
return "", err
}
return instance.InstanceIdSets.InstanceIdSet[0], nil
}
func ensureSecurityGroup(c aliclient.ECS, securityGroupName, vpcID string, log logr.Logger) (string, error) {
response, err := c.GetSecurityGroup(securityGroupName)
if err != nil {
return "", err
}
if len(response.SecurityGroups.SecurityGroup) > 0 && response.SecurityGroups.SecurityGroup[0].SecurityGroupName == securityGroupName {
log.Info("Security Group found", "security group", securityGroupName)
return response.SecurityGroups.SecurityGroup[0].SecurityGroupId, nil
}
log.Info("creating Security Group")
createResponse, err := c.CreateSecurityGroups(vpcID, securityGroupName)
if err != nil {
return "", err
}
return createResponse.SecurityGroupId, nil
}
func ensureSecurityGroupRules(c aliclient.ECS, opt *Options, shootSecurityGroupId string, bastion *extensionsv1alpha1.Bastion, securityGroupId string) error {
// ingress permission
ingressPermissions, err := ingressPermissions(bastion)
if err != nil {
return err
}
var wantedIngressRules []*ecs.AuthorizeSecurityGroupRequest
for _, ingressPermission := range ingressPermissions {
wantedIngressRules = append(wantedIngressRules, ingressAllowSSH(securityGroupId, ingressPermission))
}
currentIngressRules, err := c.DescribeSecurityGroupAttribute(describeSecurityGroupAttributeRequest(securityGroupId, "ingress"))
if err != nil {
return err
}
rulesToAddIngress, rulesToDelete := ingressRulesSymmetricDifference(wantedIngressRules, currentIngressRules.Permissions.Permission)
for _, rule := range rulesToAddIngress {
if err := c.CreateIngressRule(&rule); err != nil {
return fmt.Errorf("failed to add security group rule %s: %w", rule.Description, err)
}
}
for _, rule := range rulesToDelete {
if err := c.RevokeIngressRule(revokeSecurityGroupRequest(securityGroupId, rule.IpProtocol, rule.PortRange, rule.SourceCidrIp, rule.Ipv6SourceCidrIp)); err != nil {
return fmt.Errorf("failed to delete security group rule %s: %w", rule.Description, err)
}
}
// egress rules create
instanceResponse, err := c.GetInstances(opt.BastionInstanceName)
if err != nil {
return err
}
if len(instanceResponse.Instances.Instance) == 0 || len(instanceResponse.Instances.Instance[0].VpcAttributes.PrivateIpAddress.IpAddress) == 0 {
return errors.New("bastion instance does not have a private ip")
}
privateIP := instanceResponse.Instances.Instance[0].VpcAttributes.PrivateIpAddress.IpAddress[0]
wantedEgressRules := []*ecs.AuthorizeSecurityGroupEgressRequest{
egressAllowSSHToWorker(privateIP, securityGroupId, shootSecurityGroupId),
egressDenyAll(securityGroupId)}
currentEgressRules, err := c.DescribeSecurityGroupAttribute(describeSecurityGroupAttributeRequest(securityGroupId, "egress"))
if err != nil {
return err
}
rulesToAddEgress, rulesToDelete := egressRulesSymmetricDifference(wantedEgressRules, currentEgressRules.Permissions.Permission)
for _, rule := range rulesToAddEgress {
if err = c.CreateEgressRule(&rule); err != nil {
return fmt.Errorf("failed to add security group egress rule %s: %w", rule.Description, err)
}
}
for _, rule := range rulesToDelete {
if err = c.RevokeEgressRule(revokeSecurityGroupEgressRequest(securityGroupId, rule.IpProtocol, rule.PortRange)); err != nil {
return fmt.Errorf("failed to delete security egress group rule %s: %w", rule.Description, err)
}
}
return nil
}
func isInstanceReady(c aliclient.ECS, opt *Options) (bool, error) {
response, err := c.GetInstances(opt.BastionInstanceName)
if err != nil {
return false, err
}
if response == nil || len(response.Instances.Instance) == 0 {
return false, errors.New("bastion instance not yet created")
}
if response.Instances.Instance[0].Status == "Running" {
return true, nil
}
time.Sleep(10 * time.Second)
return false, errors.New("bastion instance not yet in Running status")
}
func ingressRulesSymmetricDifference(wantedIngressRules []*ecs.AuthorizeSecurityGroupRequest, currentRules []ecs.Permission) ([]ecs.AuthorizeSecurityGroupRequest, []ecs.Permission) {
var rulesToDelete []ecs.Permission
for _, currentRule := range currentRules {
found := false
for _, wantedRule := range wantedIngressRules {
if ingressRuleEqual(*wantedRule, currentRule) {
found = true
break
}
}
if !found {
rulesToDelete = append(rulesToDelete, currentRule)
}
}
var rulesToAdd []ecs.AuthorizeSecurityGroupRequest
for _, wantedRule := range wantedIngressRules {
found := false
for _, currentRule := range currentRules {
if ingressRuleEqual(*wantedRule, currentRule) {
found = true
break
}
}
if !found {
rulesToAdd = append(rulesToAdd, *wantedRule)
}
}
return rulesToAdd, rulesToDelete
}
func ingressRuleEqual(a ecs.AuthorizeSecurityGroupRequest, b ecs.Permission) bool {
if !equality.Semantic.DeepEqual(a.Description, b.Description) {
return false
}
if !equality.Semantic.DeepEqual(a.IpProtocol, b.IpProtocol) {
return false
}
if !equality.Semantic.DeepEqual(a.PortRange, b.PortRange) {
return false
}
if !equality.Semantic.DeepEqual(a.SourceCidrIp, b.SourceCidrIp) {
return false
}
if !equality.Semantic.DeepEqual(a.Ipv6SourceCidrIp, b.Ipv6SourceCidrIp) {
return false
}
return true
}
func egressRulesSymmetricDifference(wantedIngressRules []*ecs.AuthorizeSecurityGroupEgressRequest, currentRules []ecs.Permission) ([]ecs.AuthorizeSecurityGroupEgressRequest, []ecs.Permission) {
var rulesToDelete []ecs.Permission
for _, currentRule := range currentRules {
found := false
for _, wantedRule := range wantedIngressRules {
if egressRuleEqual(*wantedRule, currentRule) {
found = true
break
}
}
if !found {
rulesToDelete = append(rulesToDelete, currentRule)
}
}
var rulesToAdd []ecs.AuthorizeSecurityGroupEgressRequest
for _, wantedRule := range wantedIngressRules {
found := false
for _, currentRule := range currentRules {
if egressRuleEqual(*wantedRule, currentRule) {
found = true
break
}
}
if !found {
rulesToAdd = append(rulesToAdd, *wantedRule)
}
}
return rulesToAdd, rulesToDelete
}
func egressRuleEqual(a ecs.AuthorizeSecurityGroupEgressRequest, b ecs.Permission) bool {
if !equality.Semantic.DeepEqual(a.Description, b.Description) {
return false
}
if !equality.Semantic.DeepEqual(a.IpProtocol, b.IpProtocol) {
return false
}
if !equality.Semantic.DeepEqual(a.PortRange, b.PortRange) {
return false
}
if !equality.Semantic.DeepEqual(a.SourceCidrIp, b.SourceCidrIp) {
return false
}
return true
}