简介
Flannel是CoreOS团队针对Kubernetes设计的一个网络规划服务,简单来说,它的功能是让集群中的不同节点主机创建的Docker容器都具有全集群唯一的虚拟IP地址。
在默认的Docker配置中,每个节点上的Docker服务会分别负责所在节点容器的IP分配。这样导致的一个问题是,不同节点上容器可能获得相同的内外IP地址。并使这些容器之间能够之间通过IP地址相互找到,也就是相互ping通。
Flannel的设计目的就是为集群中的所有节点重新规划IP地址的使用规则,从而使得不同节点上的容器能够获得“同属一个内网”且”不重复的”IP地址,并让属于不同节点上的容器能够直接通过内网IP通信。
Flannel实质上是一种“覆盖网络(overlaynetwork)”,也就是将TCP数据包装在另一种网络包里面进行路由转发和通信,目前已经支持udp、vxlan、host-gw、aws-vpc、gce和alloc路由等数据转发方式,默认的节点间数据通信方式是UDP转发。
简单总结flanel特点
1.使集群中的不同Node主机创建的Docker容器都具有全集群唯一的虚拟IP地址。
2.建立一个覆盖网络(overlay network),通过这个覆盖网络,将数据包原封不动的传递到目标容器。覆盖网络是建立在另一个网络之上并由其基础设施支持的虚拟网络。覆盖网络通过将一个分组封装在另一个分组内来将网络服务与底层基础设施分离。在将封装的数据包转发到端点后,将其解封装。
3.创建一个新的虚拟网卡flannel0接收docker网桥的数据,通过维护路由表,对接收到的数据进行封包和转发(vxlan)。
4.etcd保证了所有node上flanned所看到的配置是一致的。同时每个node上的flanned监听etcd上的数据变化,实时感知集群中node的变化。
flannel对网络要求提出的解决办法
1.flannel利用Kubernetes API或者etcd用于存储整个集群的网络配置,根据配置记录集群使用的网段。
2.flannel在每个主机中运行flanneld作为agent,它会为所在主机从集群的网络地址空间中,获取一个小的网段subnet,本主机内所有容器的IP地址都将从中分配。
---apiVersion: policy/v1beta1kind: PodSecurityPolicymetadata:name: psp.flannel.unprivilegedannotations:seccomp.security.alpha.kubernetes.io/allowedProfileNames: docker/defaultseccomp.security.alpha.kubernetes.io/defaultProfileName: docker/defaultapparmor.security.beta.kubernetes.io/allowedProfileNames: runtime/defaultapparmor.security.beta.kubernetes.io/defaultProfileName: runtime/defaultspec:privileged: falsevolumes:- configMap- secret- emptyDir- hostPathallowedHostPaths:- pathPrefix: "/etc/cni/net.d"- pathPrefix: "/etc/kube-flannel"- pathPrefix: "/run/flannel"readOnlyRootFilesystem: false# Users and groupsrunAsUser:rule: RunAsAnysupplementalGroups:rule: RunAsAnyfsGroup:rule: RunAsAny# Privilege EscalationallowPrivilegeEscalation: falsedefaultAllowPrivilegeEscalation: false# CapabilitiesallowedCapabilities: ['NET_ADMIN', 'NET_RAW']defaultAddCapabilities: []requiredDropCapabilities: []# Host namespaceshostPID: falsehostIPC: falsehostNetwork: truehostPorts:- min: 0max: 65535# SELinuxseLinux:# SELinux is unused in CaaSPrule: 'RunAsAny'---kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1metadata:name: flannelrules:- apiGroups: ['extensions']resources: ['podsecuritypolicies']verbs: ['use']resourceNames: ['psp.flannel.unprivileged']- apiGroups:- ""resources:- podsverbs:- get- apiGroups:- ""resources:- nodesverbs:- list- watch- apiGroups:- ""resources:- nodes/statusverbs:- patch---kind: ClusterRoleBindingapiVersion: rbac.authorization.k8s.io/v1metadata:name: flannelroleRef:apiGroup: rbac.authorization.k8s.iokind: ClusterRolename: flannelsubjects:- kind: ServiceAccountname: flannelnamespace: kube-system---apiVersion: v1kind: ServiceAccountmetadata:name: flannelnamespace: kube-system---kind: ConfigMapapiVersion: v1metadata:name: kube-flannel-cfgnamespace: kube-systemlabels:tier: nodeapp: flanneldata:cni-conf.json: |{"name": "cbr0","cniVersion": "0.3.1","plugins": [{"type": "flannel","delegate": {"hairpinMode": true,"isDefaultGateway": true}},{"type": "portmap","capabilities": {"portMappings": true}}]}net-conf.json: |{"Network": "10.244.0.0/16","Backend": {"Type": "vxlan"}}---apiVersion: apps/v1kind: DaemonSetmetadata:name: kube-flannel-dsnamespace: kube-systemlabels:tier: nodeapp: flannelspec:selector:matchLabels:app: flanneltemplate:metadata:labels:tier: nodeapp: flannelspec:affinity:nodeAffinity:requiredDuringSchedulingIgnoredDuringExecution:nodeSelectorTerms:- matchExpressions:- key: kubernetes.io/osoperator: Invalues:- linuxhostNetwork: truepriorityClassName: system-node-criticaltolerations:- operator: Existseffect: NoScheduleserviceAccountName: flannelinitContainers:- name: install-cni-plugin#image: flannelcni/flannel-cni-plugin:v1.0.1 for ppc64le and mips64le (dockerhub limitations may apply)image: rancher/mirrored-flannelcni-flannel-cni-plugin:v1.0.1command:- cpargs:- -f- /flannel- /opt/cni/bin/flannelvolumeMounts:- name: cni-pluginmountPath: /opt/cni/bin- name: install-cni#image: flannelcni/flannel:v0.17.0 for ppc64le and mips64le (dockerhub limitations may apply)image: rancher/mirrored-flannelcni-flannel:v0.17.0command:- cpargs:- -f- /etc/kube-flannel/cni-conf.json- /etc/cni/net.d/10-flannel.conflistvolumeMounts:- name: cnimountPath: /etc/cni/net.d- name: flannel-cfgmountPath: /etc/kube-flannel/containers:- name: kube-flannel#image: flannelcni/flannel:v0.17.0 for ppc64le and mips64le (dockerhub limitations may apply)image: rancher/mirrored-flannelcni-flannel:v0.17.0command:- /opt/bin/flanneldargs:- --ip-masq- --kube-subnet-mgrresources:requests:cpu: "100m"memory: "50Mi"limits:cpu: "100m"memory: "50Mi"securityContext:privileged: falsecapabilities:add: ["NET_ADMIN", "NET_RAW"]env:- name: POD_NAMEvalueFrom:fieldRef:fieldPath: metadata.name- name: POD_NAMESPACEvalueFrom:fieldRef:fieldPath: metadata.namespacevolumeMounts:- name: runmountPath: /run/flannel- name: flannel-cfgmountPath: /etc/kube-flannel/- name: xtables-lockmountPath: /run/xtables.lockvolumes:- name: runhostPath:path: /run/flannel- name: cni-pluginhostPath:path: /opt/cni/bin- name: cnihostPath:path: /etc/cni/net.d- name: flannel-cfgconfigMap:name: kube-flannel-cfg- name: xtables-lockhostPath:path: /run/xtables.locktype: FileOrCreate[root@k8smaster ~]#
