本文主要是介绍Kubernetes基础(二十六)-kubernetes的eviction机制,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!
eviction,即驱赶的意思,意思是当节点出现异常时,kubernetes将有相应的机制驱赶该节点上的Pod。eviction在openstack的nova组件中也存在。
目前kubernetes中存在两种eviction机制,分别由kube-controller-manager和kubelet实现。
1 kube-controller-manager实现的eviction
kube-controller-manager主要由多个控制器构成,而eviction的功能主要由node controller这个控制器实现。
kube-controller-manager提供了以下启动参数控制eviction
- pod-eviction-timeout:即当节点宕机该事件间隔后,开始eviction机制,驱赶宕机节点上的Pod,默认为5min
- node-eviction-rate: 驱赶速率,即驱赶Node的速率,由令牌桶流控算法实现,默认为0.1,即每秒驱赶0.1个节点,注意这里不是驱赶Pod的速率,而是驱赶节点的速率。相当于每隔10s,清空一个节点
- secondary-node-eviction-rate: 二级驱赶速率,当集群中宕机节点过多时,相应的驱赶速率也降低,默认为0.01
- unhealthy-zone-threshold:不健康zone阈值,会影响什么时候开启二级驱赶速率,默认为0.55,即当该zone中节点宕机数目超过55%,而认为该zone不健康
- large-cluster-size-threshold:大集群法制,当该zone的节点多余该阈值时,则认为该zone是一个大集群。大集群节点宕机数目超过55%时,则将驱赶速率降为0.0.1,假如是小集群,则将速率直接降为0
node-controller代码主要位于pkg/controller/node目录下。
1.1 zone
为了控制eviction,kubernete将节点划分为不同的zone,主要通过给节点加label实现
- failure-domain.beta.kubernetes.io/zone
- failure-domain.beta.kubernetes.io/region
zone名称由上述的zone和region标签组合而成,两个节点zone和region分别相同,则位于同一个zone,否则不同zone。假如二者都为空,就位于default zone
zone有四种不同的状态
- stateInitial
- stateNormal
- stateFullDisruption
- statePartialDisruption
初始化状态比较好理解,假如节点刚刚加入集群,它所在的zone刚刚被发现,则该zone的状态是initial,这是一个非常短暂的时间,其余的状态,由以下函数决定:
func (nc *NodeController) ComputeZoneState(nodeReadyConditions []*v1.NodeCondition) (int, zoneState) {readyNodes := 0notReadyNodes := 0for i := range nodeReadyConditions {if nodeReadyConditions[i] != nil && nodeReadyConditions[i].Status == v1.ConditionTrue {readyNodes++} else {notReadyNodes++}} switch { case readyNodes == 0 && notReadyNodes > 0:return notReadyNodes, stateFullDisruptioncase notReadyNodes > 2 && float32(notReadyNodes)/float32(notReadyNodes+readyNodes) >= nc.unhealthyZoneThreshold:return notReadyNodes, statePartialDisruptiondefault:return notReadyNodes, stateNormal}
}
注意这里统计的某个zone下面的节点状态,而不是所有。当该zone下面ready的节点为0,而notReady节点大于0时,即认为所有节点都宕机了,所以状态为stateFullDisruption;当notReady节点大于两个,而且notReady节点占比超过unhealthyZoneThreshold,即0.55时,认为是statePartialDisruption,其他情况则认为stateNormal。
这四种状态会如何影响eviction速度呢?看如下函数:
func (nc *NodeController) setLimiterInZone(zone string, zoneSize int, state zoneState) {switch state {case stateNormal:if nc.useTaintBasedEvictions {nc.zoneNotReadyOrUnreachableTainer[zone].SwapLimiter(nc.evictionLimiterQPS)} else {nc.zonePodEvictor[zone].SwapLimiter(nc.evictionLimiterQPS)} case statePartialDisruption:if nc.useTaintBasedEvictions {nc.zoneNotReadyOrUnreachableTainer[zone].SwapLimiter(nc.enterPartialDisruptionFunc(zoneSize))} else {nc.zonePodEvictor[zone].SwapLimiter(nc.enterPartialDisruptionFunc(zoneSize))} case stateFullDisruption:if nc.useTaintBasedEvictions {nc.zoneNotReadyOrUnreachableTainer[zone].SwapLimiter(nc.enterFullDisruptionFunc(zoneSize))} else {nc.zonePodEvictor[zone].SwapLimiter(nc.enterFullDisruptionFunc(zoneSize))} }
}//其中enterPartialDisruptionFunc就是函数ReducedQPSFunc
func (nc *NodeController) ReducedQPSFunc(nodeNum int) float32 {if int32(nodeNum) > nc.largeClusterThreshold {return nc.secondaryEvictionLimiterQPS}return 0
} //而enterFullDisruptionFunc是函数HealthyQPSFunc
func (nc *NodeController) HealthyQPSFunc(nodeNum int) float32 {return nc.evictionLimiterQPS
}
即加入zone状态是normal,那么速率为0.1,假如zone状态是FullDisruption,速率也是0.1;假如zone是PartialDisruption,假如是大集群,速率为0.0.1,小集群则直接降为0。
1.2 两种不同的eviction方法
目前node controller存在两种不同的eviction方法,即通过taint或者传统方法
if nc.useTaintBasedEvictions {go wait.Until(nc.doTaintingPass, nodeEvictionPeriod, wait.NeverStop)
} else {go wait.Until(nc.doEvictionPass, nodeEvictionPeriod, wait.NeverStop)
}
其中nodeEvictionPeriod为100ms,即每隔100ms就会执行doEvictionPass或doTaintingPass。
1.2.1 传统eviction方法
zonePodEvictor类型为map[string]*RateLimitedTimedQueue,即每个zone都有一个队列,队列带了流控算法,里面存储的是unready的节点,节点上的pod需要被eviction。
func (nc *NodeController) doEvictionPass() {nc.evictorLock.Lock()defer nc.evictorLock.Unlock()for k := range nc.zonePodEvictor {nc.zonePodEvictor[k].Try(func(value TimedValue) (bool, time.Duration) {node, err := nc.nodeLister.Get(value.Value)...nodeUid, _ := value.UID.(string)remaining, err := deletePods(nc.kubeClient, nc.recorder, value.Value, nodeUid, nc.daemonSetStore)...if remaining {glog.Infof("Pods awaiting deletion due to NodeController eviction")}return true, 0}) }
}
deletePods是驱逐节点的主要函数:
func deletePods(kubeClient clientset.Interface, recorder record.EventRecorder, nodeName, nodeUID string, daemonStore extensionslisters.DaemonSetLister) (bool, error) {remaining := falseselector := fields.OneTermEqualSelector(api.PodHostField, nodeName).String()options := metav1.ListOptions{FieldSelector: selector}pods, err := kubeClient.Core().Pods(metav1.NamespaceAll).List(options)var updateErrList []error...for _, pod := range pods.Items {// Defensive check, also needed for tests.if pod.Spec.NodeName != nodeName {continue}// 设置Pod终止理由if _, err = setPodTerminationReason(kubeClient, &pod, nodeName); err != nil {if errors.IsConflict(err) {updateErrList = append(updateErrList,fmt.Errorf("update status failed for pod %q: %v", format.Pod(&pod), err))continue}}// 该Pod正在被删除,忽略if pod.DeletionGracePeriodSeconds != nil {remaining = truecontinue}// 假如该节点是又daemonset管理,则忽略_, err := daemonStore.GetPodDaemonSets(&pod)if err == nil {continue}if err := kubeClient.Core().Pods(pod.Namespace).Delete(pod.Name, nil); err != nil {return false, err}remaining = true}...return remaining, nil
}
底层其实就是delete节点上的Pod,假如Pod是由daemonset管理,则忽略,因为即使删除了,daemonset还是会在该节点上重建。
1.2.2 taint机制
taint机制还处于试验状态,默认不开启,假如要开始,则要在所有组件上设置–feature-gates TaintNodesByCondition=true
当节点状态为unready时,打上node.alpha.kubernetes.io/notReady的taint 当节点状态为unknown时,打上node.alpha.kubernetes.io/unreachable的taint
打上taint后,必然有相应的控制器去处理:
// Run starts NoExecuteTaintManager which will run in loop until `stopCh` is closed.
func (tc *NoExecuteTaintManager) Run(stopCh <-chan struct{}) { go func(stopCh <-chan struct{}) {for { item, shutdown := tc.nodeUpdateQueue.Get()if shutdown {break } nodeUpdate := item.(*nodeUpdateItem) select { case <-stopCh: break case tc.nodeUpdateChannel <- nodeUpdate:}}}(stopCh) go func(stopCh <-chan struct{}) {for {item, shutdown := tc.podUpdateQueue.Get()if shutdown {break}podUpdate := item.(*podUpdateItem)select { case <-stopCh: breakcase tc.podUpdateChannel <- podUpdate:}}}(stopCh)for { select { case <-stopCh: break case nodeUpdate := <-tc.nodeUpdateChannel:tc.handleNodeUpdate(nodeUpdate)case podUpdate := <-tc.podUpdateChannel: // If we found a Pod update we need to empty Node queue first.priority:for { select {case nodeUpdate := <-tc.nodeUpdateChannel:tc.handleNodeUpdate(nodeUpdate)default:break priority }}// After Node queue is emptied we process podUpdate.tc.handlePodUpdate(podUpdate)}}
}func deletePodHandler(c clientset.Interface, emitEventFunc func(types.NamespacedName)) func(args *WorkArgs) error {return func(args *WorkArgs) error {ns := args.NamespacedName.Namespacename := args.NamespacedName.Nameglog.V(0).Infof("NoExecuteTaintManager is deleting Pod: %v", args.NamespacedName.String())if emitEventFunc != nil { emitEventFunc(args.NamespacedName)}var err error for i := 0; i < retries; i++ {err = c.Core().Pods(ns).Delete(name, &metav1.DeleteOptions{})if err == nil { break}time.Sleep(10 * time.Millisecond)}return err}
}
本质上还是讲带eviction节点上的pod加入到删除队列上。
2 kubelet eviction机制
kube-controller-manager的eviction机制是粗粒度的,即驱赶一个节点上的所有pod,而kubelet则是细粒度的,它驱赶的是节点上的某些Pod,驱赶哪些Pod与之前讲过的Pod的Qos机制有关。
kubelet的eviction机制,只有当节点内存和磁盘资源紧张时,才会开启,他的目的就是为了回收node节点的资源。之前提过,kubelet还有oom-killer可以回收资源,那为什么还需要eviction呢?这是因为oom-killer将Pod杀掉后,假如Pod的RestartPolicy设置为Always,则kubelet隔段时间后,仍然会在该节点上启动该Pod。而kublet eviction则会将该Pod从该节点上删除。
kubelet提供了以下参数控制eviction
- eviction-hard:一系列的阈值,比如memory.available<1Gi,即当节点可用内存低于1Gi时,会立即触发一次pod eviction
- eviction-max-pod-grace-period:eviction-soft时,终止Pod的grace时间
- eviction-minimum-reclaim:表示每一次eviction必须至少回收多少资源
- eviction-pressure-transition-period:默认为5分钟,脱离pressure condition的时间,超过阈值时,节点会被设置为memory pressure或者disk pressure,然后开启pod eviction
- eviction-soft:与hard相对应,也是一系列法制,比如memory.available<1.5Gi。但它不会立即执行pod eviction,而会等待eviction-soft-grace-period时间,假如该时间过后,依然还是达到了eviction-soft,则触发一次pod eviction
- eviction-soft-grace-period:默认为90秒
2.1 核心代码
kubelet eviction的核心代码就是如下,里面的synchronize就是核心函数。
func (m *managerImpl) Start(diskInfoProvider DiskInfoProvider, podFunc ActivePodsFunc, podCleanedUpFunc PodCleanedUpFunc, nodeProvider NodeProvider, monitoringInterval time.Duration) {// start the eviction manager monitoringgo func() {for {if evictedPods := m.synchronize(diskInfoProvider, podFunc, nodeProvider); evictedPods != nil {glog.Infof("eviction manager: pods %s evicted, waiting for pod to be cleaned up", format.Pods(evictedPods))m.waitForPodsCleanup(podCleanedUpFunc, evictedPods)} else {time.Sleep(monitoringInterval)} } }()
}
2.2 何时检测触发eviction的条件
目前主要由两种机制检测触发eviction的条件
- 第一种就是定时触发,前面的synchronize位于一个for循环,其中monitoringInterval为10s,也就是每隔10s会去检测出发条件
- 通过cgroup订阅而触发,也就是假如内存低于阈值,cgroup就会通知kubelet去执行synchronize,内核层通知应用层,通过eventfd实现
if m.config.KernelMemcgNotification && !m.notifiersInitialized { glog.Infof("eviction manager attempting to integrate with kernel memcg notification api")m.notifiersInitialized = true// start soft memory notificationerr = startMemoryThresholdNotifier(m.config.Thresholds, observations, false, func(desc string) {glog.Infof("soft memory eviction threshold crossed at %s", desc)// TODO wait grace period for soft memory limitm.synchronize(diskInfoProvider, podFunc, nodeProvider)})if err != nil { glog.Warningf("eviction manager: failed to create hard memory threshold notifier: %v", err)}// start hard memory notificationerr = startMemoryThresholdNotifier(m.config.Thresholds, observations, true, func(desc string) { glog.Infof("hard memory eviction threshold crossed at %s", desc)m.synchronize(diskInfoProvider, podFunc, nodeProvider)})if err != nil {glog.Warningf("eviction manager: failed to create soft memory threshold notifier: %v", err)}}
2.3 资源的回收
kubelet的eviction主要会回收两种资源,内存和磁盘
- 磁盘回收:主要通过删除已经终止的容器和未使用的镜像
- 内存回收:主要通过终止正在运行的Pod
2.4 Qos对Eviction的影响
eviction manager会获取该节点上所有的容器,然后根据一定的算法对Pod进行排序,这里看看针对内存如何排序。
// rankMemoryPressure orders the input pods for eviction in response to memory pressure.
func rankMemoryPressure(pods []*v1.Pod, stats statsFunc) {orderedBy(qosComparator, memory(stats)).Sort(pods)
}// qosComparator compares pods by QoS (BestEffort < Burstable < Guaranteed)
func qosComparator(p1, p2 *v1.Pod) int {qosP1 := v1qos.GetPodQOS(p1)qosP2 := v1qos.GetPodQOS(p2)// its a tieif qosP1 == qosP2 {return 0} // if p1 is best effort, we know p2 is burstable or guaranteedif qosP1 == v1.PodQOSBestEffort {return -1}// we know p1 and p2 are not besteffort, so if p1 is burstable, p2 must be guaranteedif qosP1 == v1.PodQOSBurstable {if qosP2 == v1.PodQOSGuaranteed { return -1}return 1}// ok, p1 must be guaranteed.return 1
}
从qosComparator函数可以看出,Pod排列顺序为:PodQOSBestEffort < PodQOSBurstable < PodQOSGuaranteed,即首先会回收BestEffort的Pod,然后回收Burstable,最后才会回收Guranteed。
2.5 Eviction的本质
// we kill at most a single pod during each eviction intervalfor i := range activePods {pod := activePods[i] // If the pod is marked as critical and static, and support for critical pod annotations is enabled,// do not evict such pods. Static pods are not re-admitted after evictions.// https://github.com/kubernetes/kubernetes/issues/40573 has more details.if utilfeature.DefaultFeatureGate.Enabled(features.ExperimentalCriticalPodAnnotation) &&kubelettypes.IsCriticalPod(pod) && kubepod.IsStaticPod(pod) {continue }status := v1.PodStatus{ Phase: v1.PodFailed, Message: fmt.Sprintf(message, resourceToReclaim),Reason: reason, }// record that we are evicting the podm.recorder.Eventf(pod, v1.EventTypeWarning, reason, fmt.Sprintf(message, resourceToReclaim))gracePeriodOverride := int64(0)if softEviction {gracePeriodOverride = m.config.MaxPodGracePeriodSeconds}// this is a blocking call and should only return when the pod and its containers are killed.err := m.killPodFunc(pod, status, &gracePeriodOverride)if err != nil { glog.Warningf("eviction manager: error while evicting pod %s: %v", format.Pod(pod), err)} return []*v1.Pod{pod} }
每次最多回收1个Pod。
假如是hardEviction,则PodDeleteGracePeriod设置为0,即立即删除,否则设置为MaxPodGracePeriodSeconds。然后调用killPodFunc删除Pod。
需要注意的是:当kubernetes驱赶Pod的时候,kubernetes并不会重新创建Pod,假如要重新创建Pod,需要借助replicationcontroller、relicaset和deployment等机制。也就是说假如,你直接创建一个Pod,当它被kubernetes驱赶时,该Pod直接被删除了,不会重建。而利用replicationcontroller等机制,由于少了一个Pod,这些控制器就会重新创建一个Pod
这篇关于Kubernetes基础(二十六)-kubernetes的eviction机制的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!