supplement IP GC ENV for various scenarios

Signed-off-by: Icarus9913 <icaruswu66@qq.com>

charts/spiderpool/templates/deployment.yaml

@@ -161,10 +161,12 @@ spec:
  value: {{ .Values.spiderpoolController.httpPort | quote }}
  - name: SPIDERPOOL_GC_IP_ENABLED
  value: {{ .Values.ipam.gc.enabled | quote }}
- - name: SPIDERPOOL_GC_TERMINATING_POD_IP_ENABLED
- value: {{ .Values.ipam.gc.GcDeletingTimeOutPod.enabled | quote }}
+ - name: SPIDERPOOL_GC_TERMINATING_NODE_READY_POD_IP_ENABLED
+ value: {{ .Values.ipam.gc.enableGcDeletingTimeOutPodWithNodeReady | quote }}
+ - name: SPIDERPOOL_GC_TERMINATING_NODE_NOT_READY_POD_IP_ENABLED
+ value: {{ .Values.ipam.gc.enableGcDeletingTimeOutPodWithNodeNotReady | quote }}
  - name: SPIDERPOOL_GC_ADDITIONAL_GRACE_DELAY
- value: {{ .Values.ipam.gc.GcDeletingTimeOutPod.delay | quote }}
+ value: {{ .Values.ipam.gc.gcDeletingTimeOutPodDelay | quote }}
  - name: SPIDERPOOL_GC_DEFAULT_INTERVAL_DURATION
  value: {{ .Values.ipam.gc.gcAll.intervalInSecond | quote }}
  - name: SPIDERPOOL_MULTUS_CONFIG_ENABLED

charts/spiderpool/values.yaml

@@ -67,12 +67,14 @@ ipam:
  ## @param ipam.gc.gcAll.intervalInSecond the gc all interval duration
  intervalInSecond: 600
 
- GcDeletingTimeOutPod:
- ## @param ipam.gc.GcDeletingTimeOutPod.enabled enable retrieve IP for the pod who times out of deleting graceful period
- enabled: true
+ ## @param ipam.gc.enableGcDeletingTimeOutPodWithNodeReady enable reclaim IP for the pod who times out of deleting graceful period with its node ready
+ enableGcDeletingTimeOutPodWithNodeReady: true
 
- ## @param ipam.gc.GcDeletingTimeOutPod.delay the gc delay seconds after the pod times out of deleting graceful period
- delay: 0
+ ## @param ipam.gc.enableGcDeletingTimeOutPodWithNodeNotReady enable reclaim IP for the pod who times out of deleting graceful period with its node not ready
+ enableGcDeletingTimeOutPodWithNodeNotReady: true
+
+ ## @param ipam.gc.gcDeletingTimeOutPodDelay the gc delay seconds after the pod times out of deleting graceful period
+ gcDeletingTimeOutPodDelay: 0
 
 ## @section grafanaDashboard parameters
 ##

cmd/spiderpool-controller/cmd/config.go

@@ -64,7 +64,11 @@ var envInfo = []envConf{
  {"SPIDERPOOL_PYROSCOPE_PUSH_SERVER_ADDRESS", "", false, &controllerContext.Cfg.PyroscopeAddress, nil, nil},
 
  {"SPIDERPOOL_GC_IP_ENABLED", "true", true, nil, &gcIPConfig.EnableGCIP, nil},
- {"SPIDERPOOL_GC_TERMINATING_POD_IP_ENABLED", "true", true, nil, &gcIPConfig.EnableGCForTerminatingPod, nil},
+ //{"SPIDERPOOL_GC_TERMINATING_POD_IP_ENABLED", "true", true, nil, &gcIPConfig.EnableGCForTerminatingPod, nil},
+
+ {"SPIDERPOOL_GC_TERMINATING_NODE_READY_POD_IP_ENABLED", "true", true, nil, &gcIPConfig.EnableGCForTerminatingPodWithNodeReady, nil},
+ {"SPIDERPOOL_GC_TERMINATING_NODE_NOT_READY_POD_IP_ENABLED", "true", true, nil, &gcIPConfig.EnableGCForTerminatingPodWithNodeNotReady, nil},
+
  {"SPIDERPOOL_GC_IP_WORKER_NUM", "3", true, nil, nil, &gcIPConfig.ReleaseIPWorkerNum},
  {"SPIDERPOOL_GC_CHANNEL_BUFFER", "5000", true, nil, nil, &gcIPConfig.GCIPChannelBuffer},
  {"SPIDERPOOL_GC_MAX_PODENTRY_DB_CAP", "100000", true, nil, nil, &gcIPConfig.MaxPodEntryDatabaseCap},

cmd/spiderpool-controller/cmd/daemon.go

@@ -380,6 +380,7 @@ func initGCManager(ctx context.Context) {
  controllerContext.PodManager,
  controllerContext.StsManager,
  controllerContext.KubevirtManager,
+ controllerContext.NodeManager,
  controllerContext.Leader,
  )
  if nil != err {

pkg/gcmanager/gc_manager.go

@@ -17,16 +17,18 @@ import (
  "github.com/spidernet-io/spiderpool/pkg/kubevirtmanager"
  "github.com/spidernet-io/spiderpool/pkg/limiter"
  "github.com/spidernet-io/spiderpool/pkg/logutils"
+ "github.com/spidernet-io/spiderpool/pkg/nodemanager"
  "github.com/spidernet-io/spiderpool/pkg/podmanager"
  "github.com/spidernet-io/spiderpool/pkg/statefulsetmanager"
  "github.com/spidernet-io/spiderpool/pkg/workloadendpointmanager"
 )
 
 type GarbageCollectionConfig struct {
- EnableGCIP bool
- EnableGCForTerminatingPod bool
- EnableStatefulSet bool
- EnableKubevirtStaticIP bool
+ EnableGCIP bool
+ EnableGCForTerminatingPodWithNodeReady bool
+ EnableGCForTerminatingPodWithNodeNotReady bool
+ EnableStatefulSet bool
+ EnableKubevirtStaticIP bool
 
  ReleaseIPWorkerNum int
  GCIPChannelBuffer int
@@ -69,6 +71,7 @@ type SpiderGC struct {
  podMgr podmanager.PodManager
  stsMgr statefulsetmanager.StatefulSetManager
  kubevirtMgr kubevirtmanager.KubevirtManager
+ nodeMgr nodemanager.NodeManager
  leader election.SpiderLeaseElector
 
  informerFactory informers.SharedInformerFactory
@@ -81,6 +84,7 @@ func NewGCManager(clientSet *kubernetes.Clientset, config *GarbageCollectionConf
  podManager podmanager.PodManager,
  stsManager statefulsetmanager.StatefulSetManager,
  kubevirtMgr kubevirtmanager.KubevirtManager,
+ nodeMgr nodemanager.NodeManager,
  spiderControllerLeader election.SpiderLeaseElector) (GCManager, error) {
  if clientSet == nil {
  return nil, fmt.Errorf("k8s ClientSet must be specified")
@@ -121,6 +125,7 @@ func NewGCManager(clientSet *kubernetes.Clientset, config *GarbageCollectionConf
  podMgr: podManager,
  stsMgr: stsManager,
  kubevirtMgr: kubevirtMgr,
+ nodeMgr: nodeMgr,
 
  leader: spiderControllerLeader,
  gcLimiter: limiter.NewLimiter(limiter.LimiterConfig{}),

pkg/gcmanager/pod_cache.go

@@ -17,6 +17,7 @@ import (
  "github.com/spidernet-io/spiderpool/pkg/constant"
  "github.com/spidernet-io/spiderpool/pkg/lock"
  "github.com/spidernet-io/spiderpool/pkg/logutils"
+ "github.com/spidernet-io/spiderpool/pkg/nodemanager"
  "github.com/spidernet-io/spiderpool/pkg/types"
 )
 
@@ -224,9 +225,19 @@ func (s *SpiderGC) buildPodEntry(oldPod, currentPod *corev1.Pod, deleted bool) (
  }
 
  if isBuildTerminatingPodEntry {
- // disable for gc terminating pod
- if !s.gcConfig.EnableGCForTerminatingPod {
- logger.Sugar().Debugf("IP gc already turn off 'EnableGCForTerminatingPod' configuration, disacrd tracing pod '%s/%s'", currentPod.Namespace, currentPod.Name)
+ // check terminating Pod corresponding Node status
+ node, err := s.nodeMgr.GetNodeByName(ctx, currentPod.Spec.NodeName, constant.UseCache)
+ if nil != err {
+ return nil, fmt.Errorf("failed to get terminating Pod '%s/%s' corredponing Node '%s', error: %v", currentPod.Namespace, currentPod.Name, currentPod.Spec.NodeName)
+ }
+ // disable for gc terminating pod with Node Ready
+ if nodemanager.IsNodeReady(node) && !s.gcConfig.EnableGCForTerminatingPodWithNodeReady {
+ logger.Sugar().Debugf("IP GC already turn off 'EnableGCForTerminatingPodWithNodeReady' configuration, disacrd tracing pod '%s/%s'", currentPod.Namespace, currentPod.Name)
+ return nil, nil
+ }
+ // disable for gc terminating pod with Node NotReady
+ if !nodemanager.IsNodeReady(node) && !s.gcConfig.EnableGCForTerminatingPodWithNodeNotReady {
+ logger.Sugar().Debugf("IP GC already turn off 'EnableGCForTerminatingPodWithNodeNotReady' configuration, disacrd tracing pod '%s/%s'", currentPod.Namespace, currentPod.Name)
  return nil, nil
  }
 

pkg/nodemanager/utils.go

@@ -0,0 +1,18 @@
+// Copyright 2023 Authors of spidernet-io
+// SPDX-License-Identifier: Apache-2.0
+
+package nodemanager
+
+import corev1 "k8s.io/api/core/v1"
+
+func IsNodeReady(node *corev1.Node) bool {
+ var readyCondition corev1.NodeCondition
+ for _, tmpCondition := range node.Status.Conditions {
+ if tmpCondition.Type == corev1.NodeReady {
+ readyCondition = tmpCondition
+ break
+ }
+ }
+
+ return readyCondition.Status == corev1.ConditionTrue
+}

pkg/nodemanager/utils_test.go

@@ -0,0 +1,48 @@
+// Copyright 2023 Authors of spidernet-io
+// SPDX-License-Identifier: Apache-2.0
+
+package nodemanager
+
+import (
+ . "github.com/onsi/ginkgo/v2"
+ . "github.com/onsi/gomega"
+ corev1 "k8s.io/api/core/v1"
+ metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+)
+
+var _ = Describe("NodeManager utils", Label("node_manager_utils_test"), func() {
+ Describe("IsNodeReady", func() {
+ var node *corev1.Node
+ BeforeEach(func() {
+ node = &corev1.Node{
+ ObjectMeta: metav1.ObjectMeta{
+ Name: "master",
+ },
+ Spec: corev1.NodeSpec{},
+ Status: corev1.NodeStatus{
+ Conditions: []corev1.NodeCondition{
+ {
+ Type: corev1.NodeMemoryPressure,
+ Status: corev1.ConditionFalse,
+ },
+ {
+ Type: corev1.NodeReady,
+ Status: corev1.ConditionTrue,
+ },
+ },
+ },
+ }
+ })
+
+ It("Node is ready", func() {
+ isNodeReady := IsNodeReady(node)
+ Expect(isNodeReady).To(BeTrue())
+ })
+
+ It("Node is not ready", func() {
+ node.Status.Conditions[1].Status = corev1.ConditionUnknown
+ isNodeReady := IsNodeReady(node)
+ Expect(isNodeReady).To(BeFalse())
+ })
+ })
+})