Port kubernetes package to supervisor
This replaces the ad-hoc goroutine and process management
previously in the kubernetes package with a nice supervisor-based
implementation which should make it easier to understand and more
reliable. It also prevents creation of more ad-hoc launching code
for future features (like CSI & Provisioning).
Since porting SmalltownNode is rather involved I just instantiated a
new supervision tree in the Kubernetes main service and wired it
up to the old interface. Once we port SmalltownNode we can just
remove the legacy Start() method and directly call Run().
Test Plan:
Passes Bazel tests, Kubernetes functionality was manually
tested by running `bazel run //core/cmd/dbg -- kubectl run -i --image alpine:edge sh`
to verify that Kubernetes still works properly. Automated tests for this
are being worked on.
X-Origin-Diff: phab/D534
GitOrigin-RevId: 001de38eaa5c7ee661bf5db9a7c3d0125c1b6af2
diff --git a/core/internal/kubernetes/service.go b/core/internal/kubernetes/service.go
index 5e28292..ae93f4e 100644
--- a/core/internal/kubernetes/service.go
+++ b/core/internal/kubernetes/service.go
@@ -20,20 +20,17 @@
"context"
"crypto/ed25519"
"errors"
- "fmt"
"net"
+ "go.etcd.io/etcd/clientv3"
+ "go.uber.org/zap"
"google.golang.org/grpc/codes"
"google.golang.org/grpc/status"
schema "git.monogon.dev/source/nexantic.git/core/generated/api"
- "git.monogon.dev/source/nexantic.git/core/pkg/logbuffer"
-
- "go.etcd.io/etcd/clientv3"
- "go.uber.org/zap"
-
- "git.monogon.dev/source/nexantic.git/core/internal/common/service"
+ "git.monogon.dev/source/nexantic.git/core/internal/common/supervisor"
"git.monogon.dev/source/nexantic.git/core/internal/consensus"
+ "git.monogon.dev/source/nexantic.git/core/pkg/logbuffer"
)
type Config struct {
@@ -43,7 +40,6 @@
}
type Service struct {
- *service.BaseService
consensusService *consensus.Service
logger *zap.Logger
apiserverLogs *logbuffer.LogBuffer
@@ -61,7 +57,6 @@
schedulerLogs: logbuffer.New(5000, 16384),
kubeletLogs: logbuffer.New(5000, 16384),
}
- s.BaseService = service.NewBaseService("kubernetes", logger, s)
return s
}
@@ -107,65 +102,66 @@
return &schema.GetDebugKubeconfigResponse{DebugKubeconfig: string(debugKubeconfig)}, nil
}
-func (s *Service) OnStart() error {
- config := Config{
- AdvertiseAddress: net.IP{10, 0, 2, 15}, // Depends on networking
- ServiceIPRange: net.IPNet{ // TODO: Decide if configurable / final value
- IP: net.IP{192, 168, 188, 0},
- Mask: net.IPMask{0xff, 0xff, 0xff, 0x00}, // /24, but Go stores as a literal mask
- },
- ClusterNet: net.IPNet{
- IP: net.IP{192, 168, 188, 0},
- Mask: net.IPMask{0xff, 0xff, 0xfd, 0x00}, // /22
- },
- }
- consensusKV := s.getKV()
- apiserverConfig, err := getPKIApiserverConfig(consensusKV)
- if err != nil {
- return err
- }
- apiserverConfig.advertiseAddress = config.AdvertiseAddress
- apiserverConfig.serviceIPRange = config.ServiceIPRange
- controllerManagerConfig, err := getPKIControllerManagerConfig(consensusKV)
- if err != nil {
- return err
- }
- controllerManagerConfig.clusterNet = config.ClusterNet
- schedulerConfig, err := getPKISchedulerConfig(consensusKV)
- if err != nil {
- return err
- }
-
- masterKubeconfig, err := getSingle(consensusKV, "master.kubeconfig")
- if err != nil {
- return err
- }
-
- // TODO(lorenz): Once internal/node is part of the supervisor tree, these should all be supervisor runnables
- go func() {
- s.runAPIServer(context.TODO(), *apiserverConfig)
- }()
- go func() {
- s.runControllerManager(context.TODO(), *controllerManagerConfig)
- }()
- go func() {
- s.runScheduler(context.TODO(), *schedulerConfig)
- }()
-
- go func() {
- if err := s.runKubelet(context.TODO(), &KubeletSpec{}); err != nil {
- fmt.Printf("Failed to launch kubelet: %v\n", err)
- }
- }()
-
- go func() {
- go runReconciler(context.TODO(), masterKubeconfig, s.logger)
- }()
-
+func (s *Service) Start() error {
+ // TODO(lorenz): This creates a new supervision tree, it should instead attach to the root one. But for that SmalltownNode needs
+ // to be ported to supervisor.
+ supervisor.New(context.TODO(), s.logger, s.Run())
return nil
}
-func (s *Service) OnStop() error {
- // Requires advanced process management and not necessary for MVP
- return errors.New("Not implemented")
+func (s *Service) Run() supervisor.Runnable {
+ return func(ctx context.Context) error {
+ config := Config{
+ AdvertiseAddress: net.IP{10, 0, 2, 15}, // Depends on networking
+ ServiceIPRange: net.IPNet{ // TODO: Decide if configurable / final value
+ IP: net.IP{192, 168, 188, 0},
+ Mask: net.IPMask{0xff, 0xff, 0xff, 0x00}, // /24, but Go stores as a literal mask
+ },
+ ClusterNet: net.IPNet{
+ IP: net.IP{192, 168, 188, 0},
+ Mask: net.IPMask{0xff, 0xff, 0xfd, 0x00}, // /22
+ },
+ }
+ consensusKV := s.getKV()
+ apiserverConfig, err := getPKIApiserverConfig(consensusKV)
+ if err != nil {
+ return err
+ }
+ apiserverConfig.advertiseAddress = config.AdvertiseAddress
+ apiserverConfig.serviceIPRange = config.ServiceIPRange
+ controllerManagerConfig, err := getPKIControllerManagerConfig(consensusKV)
+ if err != nil {
+ return err
+ }
+ controllerManagerConfig.clusterNet = config.ClusterNet
+ schedulerConfig, err := getPKISchedulerConfig(consensusKV)
+ if err != nil {
+ return err
+ }
+
+ masterKubeconfig, err := getSingle(consensusKV, "master.kubeconfig")
+ if err != nil {
+ return err
+ }
+
+ if err := supervisor.Run(ctx, "apiserver", runAPIServer(*apiserverConfig, s.apiserverLogs)); err != nil {
+ return err
+ }
+ if err := supervisor.Run(ctx, "controller-manager", runControllerManager(*controllerManagerConfig, s.controllerManagerLogs)); err != nil {
+ return err
+ }
+ if err := supervisor.Run(ctx, "scheduler", runScheduler(*schedulerConfig, s.schedulerLogs)); err != nil {
+ return err
+ }
+
+ if err := supervisor.Run(ctx, "kubelet", runKubelet(&KubeletSpec{}, s.kubeletLogs)); err != nil {
+ return err
+ }
+ if err := supervisor.Run(ctx, "reconciler", runReconciler(masterKubeconfig)); err != nil {
+ return err
+ }
+ supervisor.Signal(ctx, supervisor.SignalHealthy)
+ supervisor.Signal(ctx, supervisor.SignalDone)
+ return nil
+ }
}