diff options
Diffstat (limited to 'docs/arm')
| -rw-r--r-- | docs/arm/data_plane_dpdk_deployment.rst | 56 | ||||
| -rw-r--r-- | docs/arm/data_plane_sriov_pf_deployment.rst | 267 | ||||
| -rw-r--r-- | docs/arm/multi_flannel_intfs_deployment.rst | 186 |
3 files changed, 270 insertions, 239 deletions
diff --git a/docs/arm/data_plane_dpdk_deployment.rst b/docs/arm/data_plane_dpdk_deployment.rst index 1f6317c..72e4011 100644 --- a/docs/arm/data_plane_dpdk_deployment.rst +++ b/docs/arm/data_plane_dpdk_deployment.rst @@ -114,33 +114,35 @@ Configuring Pod with Control Plane and Data Plane with DPDK Acceration 1, Save the below following YAML to dpdk.yaml. :: - apiVersion: v1 - kind: Pod - metadata: - name: dpdk - spec: - nodeSelector: - beta.kubernetes.io/arch: arm64 - containers: - - name: dpdk - image: younglook/dpdk:arm64 - command: [ "bash", "-c", "/usr/bin/l2fwd --huge-unlink -l 6-7 -n 4 --file-prefix=container -- -p 3" ] - stdin: true - tty: true - securityContext: - privileged: true - volumeMounts: - - mountPath: /dev/vfio - name: vfio - - mountPath: /mnt/huge - name: huge - volumes: - - name: vfio - hostPath: - path: /dev/vfio - - name: huge - hostPath: - path: /mnt/huge + .. code-block:: yaml + + apiVersion: v1 + kind: Pod + metadata: + name: dpdk + spec: + nodeSelector: + beta.kubernetes.io/arch: arm64 + containers: + - name: dpdk + image: younglook/dpdk:arm64 + command: [ "bash", "-c", "/usr/bin/l2fwd --huge-unlink -l 6-7 -n 4 --file-prefix=container -- -p 3" ] + stdin: true + tty: true + securityContext: + privileged: true + volumeMounts: + - mountPath: /dev/vfio + name: vfio + - mountPath: /mnt/huge + name: huge + volumes: + - name: vfio + hostPath: + path: /dev/vfio + - name: huge + hostPath: + path: /mnt/huge 2, Create Pod diff --git a/docs/arm/data_plane_sriov_pf_deployment.rst b/docs/arm/data_plane_sriov_pf_deployment.rst index 7cbd4d7..946d81f 100644 --- a/docs/arm/data_plane_sriov_pf_deployment.rst +++ b/docs/arm/data_plane_sriov_pf_deployment.rst @@ -16,14 +16,14 @@ This document gives a brief introduction on how to deploy SRIOV CNI with PF mode Introduction ============ -.. _sriov_cni: https://github.com/hustcat/sriov-cni -.. _Flannel: https://github.com/coreos/flannel -.. _Multus: https://github.com/Intel-Corp/multus-cni -.. _cni: https://github.com/containernetworking/cni -.. _kubeadm: https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/ -.. _k8s-crd: https://kubernetes.io/docs/concepts/api-extension/custom-resources/ -.. _arm64: https://github.com/kubernetes/website/pull/6511 -.. _files: https://github.com/kubernetes/website/pull/6511/files +.. _sriov_cni: https://github.com/hustcat/sriov-cni +.. _Flannel: https://github.com/coreos/flannel +.. _Multus: https://github.com/Intel-Corp/multus-cni +.. _cni-description: https://github.com/containernetworking/cni +.. _kubeadm: https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/ +.. _k8s-crd: https://kubernetes.io/docs/concepts/api-extension/custom-resources/ +.. _arm64: https://github.com/kubernetes/website/pull/6511 +.. _files: https://github.com/kubernetes/website/pull/6511/files As we know, in some cases we need to deploy multiple network interfaces @@ -79,20 +79,22 @@ Please make sure that rbac was added for Kubernetes cluster. here we name it as rbac.yaml: :: - apiVersion: rbac.authorization.k8s.io/v1beta1 - kind: ClusterRoleBinding - metadata: - name: fabric8-rbac - subjects: - - kind: ServiceAccount - # Reference to upper's `metadata.name` - name: default - # Reference to upper's `metadata.namespace` - namespace: default - roleRef: - kind: ClusterRole - name: cluster-admin - apiGroup: rbac.authorization.k8s.io + .. code-block:: yaml + + apiVersion: rbac.authorization.k8s.io/v1beta1 + kind: ClusterRoleBinding + metadata: + name: fabric8-rbac + subjects: + - kind: ServiceAccount + # Reference to upper's `metadata.name` + name: default + # Reference to upper's `metadata.namespace` + namespace: default + roleRef: + kind: ClusterRole + name: cluster-admin + apiGroup: rbac.authorization.k8s.io command: @@ -105,28 +107,30 @@ Please make sure that CRD was added for Kubernetes cluster. Here we name it as crdnetwork.yaml: :: - apiVersion: apiextensions.k8s.io/v1beta1 - kind: CustomResourceDefinition - metadata: - # name must match the spec fields below, and be in the form: <plural>.<group> - name: networks.kubernetes.com - spec: - # group name to use for REST API: /apis/<group>/<version> - group: kubernetes.com - # version name to use for REST API: /apis/<group>/<version> - version: v1 - # either Namespaced or Cluster - scope: Namespaced - names: - # plural name to be used in the URL: /apis/<group>/<version>/<plural> - plural: networks - # singular name to be used as an alias on the CLI and for display - singular: network - # kind is normally the CamelCased singular type. Your resource manifests use this. - kind: Network - # shortNames allow shorter string to match your resource on the CLI - shortNames: - - net + .. code-block:: yaml + + apiVersion: apiextensions.k8s.io/v1beta1 + kind: CustomResourceDefinition + metadata: + # name must match the spec fields below, and be in the form: <plural>.<group> + name: networks.kubernetes.com + spec: + # group name to use for REST API: /apis/<group>/<version> + group: kubernetes.com + # version name to use for REST API: /apis/<group>/<version> + version: v1 + # either Namespaced or Cluster + scope: Namespaced + names: + # plural name to be used in the URL: /apis/<group>/<version>/<plural> + plural: networks + # singular name to be used as an alias on the CLI and for display + singular: network + # kind is normally the CamelCased singular type. Your resource manifests use this. + kind: Network + # shortNames allow shorter string to match your resource on the CLI + shortNames: + - net command: @@ -139,19 +143,21 @@ Create flannel network as control plane. Here we name it as flannel-network.yaml: :: - apiVersion: "kubernetes.com/v1" - kind: Network - metadata: - name: flannel-conf - plugin: flannel - args: '[ - { - "masterplugin": true, - "delegate": { - "isDefaultGateway": true - } - } - ]' + .. code-block:: yaml + + apiVersion: "kubernetes.com/v1" + kind: Network + metadata: + name: flannel-conf + plugin: flannel + args: '[ + { + "masterplugin": true, + "delegate": { + "isDefaultGateway": true + } + } + ]' command: @@ -164,27 +170,29 @@ Create sriov network with PF mode as data plane. Here we name it as sriov-network.yaml: :: - apiVersion: "kubernetes.com/v1" - kind: Network - metadata: - name: sriov-conf - plugin: sriov - args: '[ - { - "master": "eth1", - "pfOnly": true, - "ipam": { - "type": "host-local", - "subnet": "192.168.123.0/24", - "rangeStart": "192.168.123.2", - "rangeEnd": "192.168.123.10", - "routes": [ - { "dst": "0.0.0.0/0" } - ], - "gateway": "192.168.123.1" - } - } - ]' + .. code-block:: yaml + + apiVersion: "kubernetes.com/v1" + kind: Network + metadata: + name: sriov-conf + plugin: sriov + args: '[ + { + "master": "eth1", + "pfOnly": true, + "ipam": { + "type": "host-local", + "subnet": "192.168.123.0/24", + "rangeStart": "192.168.123.2", + "rangeEnd": "192.168.123.10", + "routes": [ + { "dst": "0.0.0.0/0" } + ], + "gateway": "192.168.123.1" + } + } + ]' command: @@ -194,8 +202,8 @@ command: CNI Installation ================ .. _CNI: https://github.com/containernetworking/plugins -Firstly, we should deploy all CNI plugins. The build process is following: +Firstly, we should deploy all CNI plugins. The build process is following: :: git clone https://github.com/containernetworking/plugins.git @@ -219,6 +227,7 @@ we should put the Multus CNI binary to /opt/cni/bin/ where the Flannel CNI and S CNIs are put. .. _SRIOV: https://github.com/hustcat/sriov-cni + The build process of it is as: :: @@ -233,18 +242,20 @@ The following multus CNI configuration is located in /etc/cni/net.d/, here we na as multus-cni.conf: :: - { - "name": "minion-cni-network", - "type": "multus", - "kubeconfig": "/etc/kubernetes/admin.conf", - "delegates": [{ - "type": "flannel", - "masterplugin": true, - "delegate": { - "isDefaultGateway": true - } - }] - } + .. code-block:: json + + { + "name": "minion-cni-network", + "type": "multus", + "kubeconfig": "/etc/kubernetes/admin.conf", + "delegates": [{ + "type": "flannel", + "masterplugin": true, + "delegate": { + "isDefaultGateway": true + } + }] + } command: @@ -267,22 +278,24 @@ Configuring Pod with Control Plane and Data Plane In this case flannle-conf network object act as the primary network. :: - apiVersion: v1 - kind: Pod - metadata: - name: pod-sriov - annotations: - networks: '[ - { "name": "flannel-conf" }, - { "name": "sriov-conf" } - ]' - spec: # specification of the pod's contents - containers: - - name: pod-sriov - image: "busybox" - command: ["top"] - stdin: true - tty: true + .. code-block:: yaml + + apiVersion: v1 + kind: Pod + metadata: + name: pod-sriov + annotations: + networks: '[ + { "name": "flannel-conf" }, + { "name": "sriov-conf" } + ]' + spec: # specification of the pod's contents + containers: + - name: pod-sriov + image: "busybox" + command: ["top"] + stdin: true + tty: true 2, Create Pod @@ -301,25 +314,27 @@ Verifying Pod Network ===================== :: - # kubectl exec pod-sriov -- ip a - 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000 - link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 - inet 127.0.0.1/8 scope host lo - valid_lft forever preferred_lft forever - inet6 ::1/128 scope host - valid_lft forever preferred_lft forever - 3: eth0@if124: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1450 qdisc noqueue - link/ether 0a:58:0a:e9:40:2a brd ff:ff:ff:ff:ff:ff - inet 10.233.64.42/24 scope global eth0 - valid_lft forever preferred_lft forever - inet6 fe80::8e6:32ff:fed3:7645/64 scope link - valid_lft forever preferred_lft forever - 4: net0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast qlen 1000 - link/ether 52:54:00:d4:d2:e5 brd ff:ff:ff:ff:ff:ff - inet 192.168.123.2/24 scope global net0 - valid_lft forever preferred_lft forever - inet6 fe80::5054:ff:fed4:d2e5/64 scope link - valid_lft forever preferred_lft forever + .. code-block:: bash + + # kubectl exec pod-sriov -- ip a + 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000 + link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 + inet 127.0.0.1/8 scope host lo + valid_lft forever preferred_lft forever + inet6 ::1/128 scope host + valid_lft forever preferred_lft forever + 3: eth0@if124: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1450 qdisc noqueue + link/ether 0a:58:0a:e9:40:2a brd ff:ff:ff:ff:ff:ff + inet 10.233.64.42/24 scope global eth0 + valid_lft forever preferred_lft forever + inet6 fe80::8e6:32ff:fed3:7645/64 scope link + valid_lft forever preferred_lft forever + 4: net0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast qlen 1000 + link/ether 52:54:00:d4:d2:e5 brd ff:ff:ff:ff:ff:ff + inet 192.168.123.2/24 scope global net0 + valid_lft forever preferred_lft forever + inet6 fe80::5054:ff:fed4:d2e5/64 scope link + valid_lft forever preferred_lft forever Contacts ======== diff --git a/docs/arm/multi_flannel_intfs_deployment.rst b/docs/arm/multi_flannel_intfs_deployment.rst index 07c8ad7..65e643d 100644 --- a/docs/arm/multi_flannel_intfs_deployment.rst +++ b/docs/arm/multi_flannel_intfs_deployment.rst @@ -55,10 +55,10 @@ which uses Flannel as the networking backend. The related Flannel deployment fil image to start the Flannel service. .. image:: images/multi_flannel_intfs.PNG - :alt: 2 Flannel interfaces deployment scenario - :figclass: align-center + :width: 800px + :alt: 2 Flannel interfaces deployment scenario - Fig 1. Multiple Flannel interfaces deployment architecture +Fig 1. Multiple Flannel interfaces deployment architecture .. _Etcd: https://coreos.com/etcd/ @@ -84,7 +84,7 @@ kube-flannel.yml. Here we give a revised version of this yaml file to start 2 Fl .. include:: files/kube-2flannels.yml :literal: - kube-2flannels.yml +kube-2flannels.yml ConfigMap Added @@ -94,14 +94,16 @@ To start the 2nd Flannel container process, we add a new ConfigMap named kube-fl includes a new net-conf.json from the 1st: :: - net-conf.json: | - { - "Network": "10.3.0.0/16", - "Backend": { - "Type": "udp", - "Port": 8286 + .. code-block:: json + + net-conf.json: | + { + "Network": "10.3.0.0/16", + "Backend": { + "Type": "udp", + "Port": 8286 + } } - } 2nd Flannel Container Added @@ -112,20 +114,24 @@ The default Flanneld's UDP listen port is 8285, we set the 2nd Flanneld to liste For the 2nd Flannel container, we use the command as: :: - - name: kube-flannel2 - image: quay.io/coreos/flannel:v0.8.0-arm64 - command: [ "/opt/bin/flanneld", "--ip-masq", "--kube-subnet-mgr", "--subnet-file=/run/flannel/subnet2.env" ] + .. code-block:: yaml + + - name: kube-flannel2 + image: quay.io/coreos/flannel:v0.8.0-arm64 + command: [ "/opt/bin/flanneld", "--ip-masq", "--kube-subnet-mgr", "--subnet-file=/run/flannel/subnet2.env" ] which outputs the subnet file to /run/flannel/subnet2.env for the 2nd Flannel CNI to use. And mount the 2nd Flannel ConfigMap to /etc/kube-flannel/ for the 2nd Flanneld container process: :: - volumeMounts: - - name: run - mountPath: /run - - name: flannel-cfg2 - mountPath: /etc/kube-flannel/ + .. code-block:: yaml + + volumeMounts: + - name: run + mountPath: /run + - name: flannel-cfg2 + mountPath: /etc/kube-flannel/ CNI Configuration @@ -148,33 +154,35 @@ The following CNI configuration sample for 2 Flannel interfaces is located in /e as 10-2flannels.conf: :: - { - "name": "flannel-networks", - "type": "multus", - "delegates": [ - { - "type": "flannel", - "name": "flannel.2", - "subnetFile": "/run/flannel/subnet2.env", - "dataDir": "/var/lib/cni/flannel/2", - "delegate": { - "bridge": "kbr1", - "isDefaultGateway": false - } - }, - { - "type": "flannel", - "name": "flannel.1", - "subnetFile": "/run/flannel/subnet.env", - "dataDir": "/var/lib/cni/flannel", - "masterplugin": true, - "delegate": { - "bridge": "kbr0", - "isDefaultGateway": true - } - } - ] - } + .. code-block:: json + + { + "name": "flannel-networks", + "type": "multus", + "delegates": [ + { + "type": "flannel", + "name": "flannel.2", + "subnetFile": "/run/flannel/subnet2.env", + "dataDir": "/var/lib/cni/flannel/2", + "delegate": { + "bridge": "kbr1", + "isDefaultGateway": false + } + }, + { + "type": "flannel", + "name": "flannel.1", + "subnetFile": "/run/flannel/subnet.env", + "dataDir": "/var/lib/cni/flannel", + "masterplugin": true, + "delegate": { + "bridge": "kbr0", + "isDefaultGateway": true + } + } + ] + } For the 2nd Flannel CNI, it will use the subnet file /run/flannel/subnet2.env instead of the default /run/flannel/subnet.env, which is generated by the 2nd Flanneld process, and the subnet data would be output to the directory: @@ -228,32 +236,36 @@ kube-flannel.yml. For Flanneld to use the etcd backend, we could change the cont backend: :: - ... - containers: - - name: kube-flannel - image: quay.io/coreos/flannel:v0.8.0-arm64 - command: [ "/opt/bin/flanneld", "--ip-masq", "--etcd-endpoints=http://ETCD_CLUSTER_IP1:2379", "--etcd-prefix=/coreos.com/network" ] - securityContext: - privileged: true - env: - - name: POD_NAME - valueFrom: - fieldRef: - fieldPath: metadata.name - - name: POD_NAMESPACE - valueFrom: - fieldRef: - fieldPath: metadata.namespace - volumeMounts: - - name: run - mountPath: /run - - name: flannel-cfg - mountPath: /etc/kube-flannel/ + .. code-block:: yaml + + ... + containers: + - name: kube-flannel + image: quay.io/coreos/flannel:v0.8.0-arm64 + command: [ "/opt/bin/flanneld", "--ip-masq", "--etcd-endpoints=http://ETCD_CLUSTER_IP1:2379", "--etcd-prefix=/coreos.com/network" ] + securityContext: + privileged: true + env: + - name: POD_NAME + valueFrom: + fieldRef: + fieldPath: metadata.name + - name: POD_NAMESPACE + valueFrom: + fieldRef: + fieldPath: metadata.namespace + volumeMounts: + - name: run + mountPath: /run + - name: flannel-cfg + mountPath: /etc/kube-flannel/ Here as we don't use the "--kube-subnet-mgr" option, the last 2 lines of :: - - name: flannel-cfg + .. code-block:: yaml + + - name: flannel-cfg mountPath: /etc/kube-flannel/ can be ignored. @@ -262,24 +274,26 @@ To start the 2nd Flanneld process, we can add the 2nd Flanneld container section the 1st Flanneld container: :: - containers: - - name: kube-flannel2 - image: quay.io/coreos/flannel:v0.8.0-arm64 - command: [ "/opt/bin/flanneld", "--ip-masq", "--etcd-endpoints=http://ETCD_CLUSTER_IP1:2379", "--etcd-prefix=/coreos.com/network2", "--subnet-file=/run/flannel/subnet2.env" ] - securityContext: - privileged: true - env: - - name: POD_NAME - valueFrom: - fieldRef: - fieldPath: metadata.name - - name: POD_NAMESPACE - valueFrom: - fieldRef: - fieldPath: metadata.namespace - volumeMounts: - - name: run - mountPath: /run + .. code-block:: yaml + + containers: + - name: kube-flannel2 + image: quay.io/coreos/flannel:v0.8.0-arm64 + command: [ "/opt/bin/flanneld", "--ip-masq", "--etcd-endpoints=http://ETCD_CLUSTER_IP1:2379", "--etcd-prefix=/coreos.com/network2", "--subnet-file=/run/flannel/subnet2.env" ] + securityContext: + privileged: true + env: + - name: POD_NAME + valueFrom: + fieldRef: + fieldPath: metadata.name + - name: POD_NAMESPACE + valueFrom: + fieldRef: + fieldPath: metadata.namespace + volumeMounts: + - name: run + mountPath: /run The option "-subnet-file" for the 2nd Flanneld is to output a subnet file for the 2nd Flannel subnet configuration of the Flannel CNI which is called by Multus CNI. |