kube-controller-manager最佳配置
kube-controller-manager最佳配置
WaltonWang 发表于9个月前
kube-controller-manager最佳配置
  • 发表于 9个月前
  • 阅读 119
  • 收藏 1
  • 点赞 0
  • 评论 0

新睿云服务器60天免费使用,快来体验!>>>   

kubernetes version: 1.6.2

Kube-controller-manager Configuration

kube-controller-manager一共有72个flag配置,具体如下:

flagvaluedefault valuecomments
--address ip0.0.0.0The IP address to serve on (set to 0.0.0.0 for all interfaces) (default 0.0.0.0)
--allocate-node-cidrsShould CIDRs for Pods be allocated and set on the cloud provider.
--attach-detach-reconcile-sync-period duration1m0sThe reconciler sync wait time between volume attach detach. This duration must be larger than one second, and increasing this value from the default may allow for volumes to be mismatched with pods. (default 1m0s)
--azure-container-registry-config stringPath to the file container Azure container registry configuration information.
--cloud-config stringThe path to the cloud provider configuration file. Empty string for no configuration file.
--cloud-provider stringThe provider for cloud services. Empty string for no provider.
--cluster-cidr stringCIDR Range for Pods in cluster.
--cluster-name stringkubernetesThe instance prefix for the cluster (default "kubernetes")
--cluster-signing-cert-file string/etc/kubernetes/ca/ca.pemFilename containing a PEM-encoded X509 CA certificate used to issue cluster-scoped certificates (default "/etc/kubernetes/ca/ca.pem")
--cluster-signing-key-file string/etc/kubernetes/ca/ca.keyFilename containing a PEM-encoded RSA or ECDSA private key used to sign cluster-scoped certificates (default "/etc/kubernetes/ca/ca.key")
--concurrent-deployment-syncs int325The number of deployment objects that are allowed to sync concurrently. Larger number = more responsive deployments, but more CPU (and network) load (default 5)
--concurrent-endpoint-syncs int325The number of endpoint syncing operations that will be done concurrently. Larger number = faster endpoint updating, but more CPU (and network) load (default 5)
--concurrent-gc-syncs int3220The number of garbage collector workers that are allowed to sync concurrently. (default 20)
--concurrent-namespace-syncs int322The number of namespace objects that are allowed to sync concurrently. Larger number = more responsive namespace termination, but more CPU (and network) load (default 2)
--concurrent-replicaset-syncs int325The number of replica sets that are allowed to sync concurrently. Larger number = more responsive replica management, but more CPU (and network) load (default 5)
--concurrent-resource-quota-syncs int325The number of resource quotas that are allowed to sync concurrently. Larger number = more responsive quota management, but more CPU (and network) load (default 5)
--concurrent-service-syncs int321The number of services that are allowed to sync concurrently. Larger number = more responsive service management, but more CPU (and network) load (default 1)
--concurrent-serviceaccount-token-syncs int325The number of service account token objects that are allowed to sync concurrently. Larger number = more responsive token generation, but more CPU (and network) load (default 5)
--concurrent-rc-syncs int325The number of replication controllers that are allowed to sync concurrently. Larger number = more responsive replica management, but more CPU (and network) load (default 5)
--configure-cloud-routestrueShould CIDRs allocated by allocate-node-cidrs be configured on the cloud provider. (default true)
--contention-profilingEnable lock contention profiling, if profiling is enabled
--controller-start-interval durationInterval between starting controller managers.
--controllers stringSlice*A list of controllers to enable. '*' enables all on-by-default controllers, 'foo' enables the controller named 'foo', '-foo' disables the controller named 'foo'. <br/> All controllers: attachdetach, bootstrapsigner, certificatesigningrequests, cronjob, daemonset, deployment, disruption, endpoint, garbagecollector, horizontalpodautoscaling, job, namespace, node, persistentvolume-binder, podgc, replicaset, replicationcontroller, resourcequota, route, service, serviceaccount, serviceaccount-token, statefuleset, tokencleaner, ttl <br/> Disabled-by-default controllers: bootstrapsigner, tokencleaner (default [*])
--deployment-controller-sync-period duration30sPeriod for syncing the deployments. (default 30s)
--disable-attach-detach-reconcile-syncDisable volume attach detach reconciler sync. Disabling this may cause volumes to be mismatched with pods. Use wisely.
--enable-dynamic-provisioningtrueEnable dynamic provisioning for environments that support it. (default true)
--enable-garbage-collectortrueEnables the generic garbage collector. MUST be synced with the corresponding flag of the kube-apiserver. (default true)
--enable-hostpath-provisionerEnable HostPath PV provisioning when running without a cloud provider. This allows testing and development of provisioning features. HostPath provisioning is not supported in any way, won't work in a multi-node cluster, and should not be used for anything other than testing or development.
--enable-taint-managertrueWARNING: Beta feature. If set to true, enables NoExecute Taints and will evict all not-tolerating Pods running on Nodes tainted with these kinds of Taints. (default true)
--feature-gates mapStringBoolA set of key=value pairs that describe feature gates for alpha/experimental features. Options are: <br/> Accelerators=true
--flex-volume-plugin-dir string/usr/libexec/kubernetes/kubelet-plugins/volume/exec/Full path of the directory in which the flex volume plugin should search for additional third party volume plugins. (default "/usr/libexec/kubernetes/kubelet-plugins/volume/exec/")
--google-json-key stringThe Google Cloud Platform Service Account JSON Key to use for authentication.
--horizontal-pod-autoscaler-sync-period duration30sThe period for syncing the number of pods in horizontal pod autoscaler. (default 30s)
--horizontal-pod-autoscaler-use-rest-clientsWARNING: alpha feature. If set to true, causes the horizontal pod autoscaler controller to use REST clients through the kube-aggregator, instead of using the legacy metrics client through the API server proxy. This is required for custom metrics support in the horizonal pod autoscaler.
--insecure-experimental-approve-all-kubelet-csrs-for-group stringThe group for which the controller-manager will auto approve all CSRs for kubelet client certificates.
--kube-api-burst int3230Burst to use while talking with kubernetes apiserver (default 30)
--kube-api-content-type stringapplication/vnd.kubernetes.protobufContent type of requests sent to apiserver. (default "application/vnd.kubernetes.protobuf")
--kube-api-qps float3220QPS to use while talking with kubernetes apiserver (default 20)
--kubeconfig stringPath to kubeconfig file with authorization and master location information.
--large-cluster-size-threshold int3250Number of nodes from which NodeController treats the cluster as large for the eviction logic purposes. --secondary-node-eviction-rate is implicitly overridden to 0 for clusters this size or smaller. (default 50)
--leader-electtrueStart a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability. (default true)
--leader-elect-lease-duration duration15sThe duration that non-leader candidates will wait after observing a leadership renewal until attempting to acquire leadership of a led but unrenewed leader slot. This is effectively the maximum duration that a leader can be stopped before it is replaced by another candidate. This is only applicable if leader election is enabled. (default 15s)
--leader-elect-renew-deadline duration10sThe interval between attempts by the acting master to renew a leadership slot before it stops leading. This must be less than or equal to the lease duration. This is only applicable if leader election is enabled. (default 10s)
--leader-elect-retry-period duration2sThe duration the clients should wait between attempting acquisition and renewal of a leadership. This is only applicable if leader election is enabled. (default 2s)
--master stringThe address of the Kubernetes API server (overrides any value in kubeconfig)
--min-resync-period duration12h0m0The resync period in reflectors will be random between MinResyncPeriod and 2*MinResyncPeriod (default 12h0m0s)
--namespace-sync-period duration5m0sThe period for syncing namespace life-cycle updates (default 5m0s)
--node-cidr-mask-size int3224Mask size for node cidr in cluster. (default 24)
--node-eviction-rate float320.1Number of nodes per second on which pods are deleted in case of node failure when a zone is healthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters. (default 0.1)
--node-monitor-grace-period duration40sAmount of time which we allow running Node to be unresponsive before marking it unhealthy. Must be N times more than kubelet's nodeStatusUpdateFrequency, where N means number of retries allowed for kubelet to post node status. (default 40s)
--node-monitor-period duration5sThe period for syncing NodeStatus in NodeController. (default 5s)
--node-startup-grace-period duration1m0sAmount of time which we allow starting Node to be unresponsive before marking it unhealthy. (default 1m0s)
--pod-eviction-timeout duration5m0sThe grace period for deleting pods on failed nodes. (default 5m0s)
--port int3210252The port that the controller-manager's http service runs on (default 10252)
--profilingtrueEnable profiling via web interface host:port/debug/pprof/ (default true)
--pv-recycler-increment-timeout-nfs int3230the increment of time added per Gi to ActiveDeadlineSeconds for an NFS scrubber pod (default 30)
--pv-recycler-minimum-timeout-hostpath int3260The minimum ActiveDeadlineSeconds to use for a HostPath Recycler pod. This is for development and testing only and will not work in a multi-node cluster. (default 60)
--pv-recycler-minimum-timeout-nfs int32300The minimum ActiveDeadlineSeconds to use for an NFS Recycler pod (default 300)
--pv-recycler-pod-template-filepath-hostpath stringThe file path to a pod definition used as a template for HostPath persistent volume recycling. This is for development and testing only and will not work in a multi-node cluster.
--pv-recycler-pod-template-filepath-nfs stringThe file path to a pod definition used as a template for NFS persistent volume recycling
--pv-recycler-timeout-increment-hostpath int3230the increment of time added per Gi to ActiveDeadlineSeconds for a HostPath scrubber pod. This is for development and testing only and will not work in a multi-node cluster. (default 30)
--pvclaimbinder-sync-period duration15sThe period for syncing persistent volumes and persistent volume claims (default 15s)
--resource-quota-sync-period duration5m0sThe period for syncing quota usage status in the system (default 5m0s)
--root-ca-file stringIf set, this root certificate authority will be included in service account's token secret. This must be a valid PEM-encoded CA bundle.
--route-reconciliation-period duration10sThe period for reconciling routes created for Nodes by cloud provider. (default 10s)
--secondary-node-eviction-rate float320.01Number of nodes per second on which pods are deleted in case of node failure when a zone is unhealthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters. This value is implicitly overridden to 0 if the cluster size is smaller than --large-cluster-size-threshold. (default 0.01)
--service-account-private-key-file stringFilename containing a PEM-encoded private RSA or ECDSA key used to sign service account tokens.
--service-cluster-ip-range stringCIDR Range for Services in cluster.
--service-sync-period duration5m0sThe period for syncing services with their external load balancers (default 5m0s)
--terminated-pod-gc-threshold int3212500Number of terminated pods that can exist before the terminated pod garbage collector starts deleting terminated pods. If <= 0, the terminated pod garbage collector is disabled. (default 12500)
--unhealthy-zone-threshold float320.55Fraction of Nodes in a zone which needs to be not Ready (minimum 3) for zone to be treated as unhealthy. (default 0.55)
--use-service-account-credentialsIf true, use individual service account credentials for each controller.

kube-controller-manager Configuration We Should First Care About

在72个flag中,我梳理出了我们私有云环境需要注意或配置的项,如下表所示。

flagvaluedefault valuecomments
--address ip0.0.0.0The IP address to serve on (set to 0.0.0.0 for all interfaces) (default 0.0.0.0)
--attach-detach-reconcile-sync-period duration1m0sThe reconciler sync wait time between volume attach detach. This duration must be larger than one second, and increasing this value from the default may allow for volumes to be mismatched with pods. (default 1m0s)
--cluster-cidr stringCIDR Range for Pods in cluster.
--cluster-name stringkubernetesThe instance prefix for the cluster (default "kubernetes")
--cluster-signing-cert-file string/etc/kubernetes/ca/ca.pemFilename containing a PEM-encoded X509 CA certificate used to issue cluster-scoped certificates (default "/etc/kubernetes/ca/ca.pem")
--cluster-signing-key-file string/etc/kubernetes/ca/ca.keyFilename containing a PEM-encoded RSA or ECDSA private key used to sign cluster-scoped certificates (default "/etc/kubernetes/ca/ca.key")
--contention-profilingEnable lock contention profiling, if profiling is enabled
--enable-garbage-collectortrueEnables the generic garbage collector. MUST be synced with the corresponding flag of the kube-apiserver. (default true)
--flex-volume-plugin-dir string/usr/libexec/kubernetes/kubelet-plugins/volume/exec/Full path of the directory in which the flex volume plugin should search for additional third party volume plugins. (default "/usr/libexec/kubernetes/kubelet-plugins/volume/exec/")
--kubeconfig stringPath to kubeconfig file with authorization and master location information.
--large-cluster-size-threshold int3250Number of nodes from which NodeController treats the cluster as large for the eviction logic purposes. --secondary-node-eviction-rate is implicitly overridden to 0 for clusters this size or smaller. (default 50)
--master stringThe address of the Kubernetes API server (overrides any value in kubeconfig)
--node-cidr-mask-size int3224Mask size for node cidr in cluster. (default 24)
--node-eviction-rate float320.1Number of nodes per second on which pods are deleted in case of node failure when a zone is healthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters. (default 0.1)
--node-monitor-grace-period duration40sAmount of time which we allow running Node to be unresponsive before marking it unhealthy. Must be N times more than kubelet's nodeStatusUpdateFrequency, where N means number of retries allowed for kubelet to post node status. (default 40s)
--node-startup-grace-period duration1m0s ?Amount of time which we allow starting Node to be unresponsive before marking it unhealthy. (default 1m0s)
--pod-eviction-timeout duration5m0sThe grace period for deleting pods on failed nodes. (default 5m0s)
--port int3210252The port that the controller-manager's http service runs on (default 10252)
--resource-quota-sync-period duration5m0sThe period for syncing quota usage status in the system (default 5m0s)
--root-ca-file stringIf set, this root certificate authority will be included in service account's token secret. This must be a valid PEM-encoded CA bundle.
--secondary-node-eviction-rate float320.01Number of nodes per second on which pods are deleted in case of node failure when a zone is unhealthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters. This value is implicitly overridden to 0 if the cluster size is smaller than --large-cluster-size-threshold. (default 0.01)
--service-account-private-key-file stringFilename containing a PEM-encoded private RSA or ECDSA key used to sign service account tokens.
--service-cluster-ip-range stringCIDR Range for Services in cluster.
--terminated-pod-gc-threshold int3212500Number of terminated pods that can exist before the terminated pod garbage collector starts deleting terminated pods. If <= 0, the terminated pod garbage collector is disabled. (default 12500)
--unhealthy-zone-threshold float320.55Fraction of Nodes in a zone which needs to be not Ready (minimum 3) for zone to be treated as unhealthy. (default 0.55)

更多关于kubernetes的深入文章,请看我csdn或者oschina的博客主页。

标签: Kubernetes config
  • 打赏
  • 点赞
  • 收藏
  • 分享
共有 人打赏支持
粉丝 115
博文 77
码字总数 145830
×
WaltonWang
如果觉得我的文章对您有用,请随意打赏。您的支持将鼓励我继续创作!
* 金额(元)
¥1 ¥5 ¥10 ¥20 其他金额
打赏人
留言
* 支付类型
微信扫码支付
打赏金额:
已支付成功
打赏金额: