【kubernetes/k8s概念】kube-scheduler启动参数
kubernetes 1.12.1版本Desc The Kubernetes scheduler is a policy-rich, topology-aware, workload-specific function that significantly impacts availability, performance, and capacity. The scheduler...
kubernetes 1.12.1版本
Desc
The Kubernetes scheduler is a policy-rich, topology-aware, workload-specific function that significantly impacts availability, performance, and capacity. The scheduler needs to take into account individual and collective resource requirements, quality of service requirements, hardware/software/policy constraints, affinity and anti-affinity specifications, data locality, inter-workload interference, deadlines, and so on. Workload-specific requirements will be exposed through the API as necessary.
kube-scheduler通过调度算法为待调度Pod列表上的每一个Pod从Node列表中选择一个最合适的Node。
kube-scheduler当前提供的默认调度流程分为两步
- 预选调度过程:遍历所有目标Node,筛选出符合要求的候选节点
- 确定最优节点:在第一步的基础上,采用优选策略计算出每个候选节点的分数,分数最高者胜出
/opt/k8s/bin/kube-scheduler --address=0.0.0.0 --master=http://127.0.0.1:8080 --leader-elect=true --v=2
kube-scheduler启动参数详解
Usage:
kube-scheduler [flags]
--address string | 监听主机IP地址,0.0.0.0监听主机所有主机接口 |
--algorithm-provider string | 设置调度算法,ClusterAutoscalerProvider | DefaultProvider,默认为DefaultProvider |
--alsologtostderr | 设置true则日志输出到stderr,也输出到日志文件 |
--config string | 配置文件的路径 |
--kube-api-burst int32 | 发送到kube-apiserver每秒请求量 (default 100) |
--kube-api-content-type string | 发送到kube-apiserver请求内容类型(default "application/vnd.kubernetes.protobuf") |
--kube-api-qps float32 | 与kube-apiserver通信的qps(default 50) |
--kubeconfig string | kubeconfig配置文件路径,该文件包括master认证以及master信息 |
--leader-elect | 多个master情况设置为true保证高可用,进行leader选举 |
--leader-elect-lease-duration duration | 当leader-elect设置为true生效,选举过程中非leader候选等待选举的时间间隔(default 15s) |
--leader-elect-renew-deadline duration | leader选举过程中在停止leading,再次renew时间间隔,小于或者等于leader-elect-lease-duration duration,也是leader-elect设置为true生效(default 10s) |
--leader-elect-retry-period duration | 当leader-elect设置为true生效,获取leader或者重新选举的等待间隔(default 2s) |
--lock-object-name string | 定义lock对象名字(default "kube-scheduler") |
--lock-object-namespace string | 定义lock对象的namespace(default "kube-system") |
--log-backtrace-at traceLocation | 记录日志到file:行号时打印一次stack trace(default :0) |
--log-dir string | 记录log的目录 |
--log-flush-frequency duration | flush log的时间间隔(default 5s) |
--logtostderr | 写log到stderr(default true) |
--master string | master的地址,会覆盖kubeconfig中的 |
--port int | 没有认证鉴权的不安全端口(default 10251) |
--profiling | 开启性能分析,通过host:port/debug/pprof/查看 |
--scheduler-name string | 调度器名,由于哪些pod被调度器进行处理,根据pod的spec.schedulerName(default "default-scheduler") |
更多推荐
所有评论(0)