06-4.部署高可用 kube-scheduler 集群

06-4.部署高可用 kube-scheduler 集群

本文档介绍部署高可用 kube-scheduler 集群的步骤。

该集群包含 3 个节点，启动后将通过竞争选举机制产生一个 leader 节点，其它节点为阻塞状态。当 leader 节点不可用后，剩余节点将再次进行选举产生新的 leader 节点，从而保证服务的可用性。

为保证通信安全，本文档先生成 x509 证书和私钥，kube-scheduler 在如下两种情况下使用该证书：

与 kube-apiserver 的安全端口通信;
在安全端口(https，10251) 输出 prometheus 格式的 metrics；

注意：如果没有特殊指明，本文档的所有操作均在 m7-autocv-gpu01 节点上执行，然后远程分发文件和执行命令。

准备工作

下载最新版本的二进制文件、安装和配置 flanneld 参考：06-1.部署master节点.md

创建 kube-scheduler 证书和私钥

创建证书签名请求：

cd /opt/k8s/work
cat > kube-scheduler-csr.json <<EOF
{
    "CN": "system:kube-scheduler",
    "hosts": [
      "127.0.0.1",
      "172.27.128.148",
      "172.27.128.149",
      "172.27.128.150"
    ],
    "key": {
        "algo": "rsa",
        "size": 2048
    },
    "names": [
      {
        "C": "CN",
        "ST": "BeiJing",
        "L": "BeiJing",
        "O": "system:kube-scheduler",
        "OU": "4Paradigm"
      }
    ]
}
EOF

hosts 列表包含所有 kube-scheduler 节点 IP；
CN 为 system:kube-scheduler、O 为 system:kube-scheduler，kubernetes 内置的 ClusterRoleBindings system:kube-scheduler 将赋予 kube-scheduler 工作所需的权限。

生成证书和私钥：

cd /opt/k8s/work
cfssl gencert -ca=/opt/k8s/work/ca.pem \
  -ca-key=/opt/k8s/work/ca-key.pem \
  -config=/opt/k8s/work/ca-config.json \
  -profile=kubernetes kube-scheduler-csr.json | cfssljson -bare kube-scheduler
ls kube-scheduler*pem

创建和分发 kubeconfig 文件

kubeconfig 文件包含访问 apiserver 的所有信息，如 apiserver 地址、CA 证书和自身使用的证书；

cd /opt/k8s/work
source /opt/k8s/bin/environment.sh
kubectl config set-cluster kubernetes \
  --certificate-authority=/opt/k8s/work/ca.pem \
  --embed-certs=true \
  --server=${KUBE_APISERVER} \
  --kubeconfig=kube-scheduler.kubeconfig
kubectl config set-credentials system:kube-scheduler \
  --client-certificate=kube-scheduler.pem \
  --client-key=kube-scheduler-key.pem \
  --embed-certs=true \
  --kubeconfig=kube-scheduler.kubeconfig
kubectl config set-context system:kube-scheduler \
  --cluster=kubernetes \
  --user=system:kube-scheduler \
  --kubeconfig=kube-scheduler.kubeconfig
kubectl config use-context system:kube-scheduler --kubeconfig=kube-scheduler.kubeconfig

上一步创建的证书、私钥以及 kube-apiserver 地址被写入到 kubeconfig 文件中；

分发 kubeconfig 到所有 master 节点：

cd /opt/k8s/work
source /opt/k8s/bin/environment.sh
for node_ip in ${NODE_IPS[@]}
  do
    echo ">>> ${node_ip}"
    scp kube-scheduler.kubeconfig root@${node_ip}:/etc/kubernetes/
  done

创建 kube-scheduler 配置文件

cd /opt/k8s/work
cat <<EOF | sudo tee kube-scheduler.yaml
apiVersion: kubescheduler.config.k8s.io/v1alpha1
kind: KubeSchedulerConfiguration
clientConnection:
  kubeconfig: "/etc/kubernetes/kube-scheduler.kubeconfig"
leaderElection:
  leaderElect: true
EOF

--kubeconfig：指定 kubeconfig 文件路径，kube-scheduler 使用它连接和验证 kube-apiserver；
--leader-elect=true：集群运行模式，启用选举功能；被选为 leader 的节点负责处理工作，其它节点为阻塞状态；

分发 kube-scheduler 配置文件到所有 master 节点：

cd /opt/k8s/work
source /opt/k8s/bin/environment.sh
for node_ip in ${NODE_IPS[@]}
  do
    echo ">>> ${node_ip}"
    scp kube-scheduler.yaml root@${node_ip}:/etc/kubernetes/
  done

创建和分发 kube-scheduler systemd unit 文件

cd /opt/k8s/work
cat > kube-scheduler.service <<EOF
[Unit]
Description=Kubernetes Scheduler
Documentation=https://github.com/GoogleCloudPlatform/kubernetes
[Service]
WorkingDirectory=${K8S_DIR}/kube-scheduler
ExecStart=/opt/k8s/bin/kube-scheduler \\
  --config=/etc/kubernetes/kube-scheduler.yaml \\
  --address=127.0.0.1 \\
  --kube-api-qps=100 \\
  --logtostderr=true \\
  --v=2
Restart=always
RestartSec=5
StartLimitInterval=0
[Install]
WantedBy=multi-user.target
EOF

--address：在 127.0.0.1:10251 端口接收 http /metrics 请求；kube-scheduler 目前还不支持接收 https 请求；

完整 unit 见 kube-scheduler.service。

分发 systemd unit 文件到所有 master 节点：

cd /opt/k8s/work
source /opt/k8s/bin/environment.sh
for node_ip in ${NODE_IPS[@]}
  do
    echo ">>> ${node_ip}"
    scp kube-scheduler.service root@${node_ip}:/etc/systemd/system/
  done

启动 kube-scheduler 服务

source /opt/k8s/bin/environment.sh
for node_ip in ${NODE_IPS[@]}
  do
    echo ">>> ${node_ip}"
    ssh root@${node_ip} "mkdir -p ${K8S_DIR}/kube-scheduler"
    ssh root@${node_ip} "systemctl daemon-reload && systemctl enable kube-scheduler && systemctl restart kube-scheduler"
  done

检查服务运行状态

source /opt/k8s/bin/environment.sh
for node_ip in ${NODE_IPS[@]}
  do
    echo ">>> ${node_ip}"
    ssh root@${node_ip} "systemctl status kube-scheduler|grep Active"
  done

确保状态为 active (running)，否则查看日志，确认原因：

$ journalctl -u kube-scheduler

查看输出的 metric

注意：以下命令在 kube-scheduler 节点上执行。

kube-scheduler 监听 10251 端口，接收 http 请求：

$ sudo netstat -lnpt|grep kube-sche
tcp        0      0 127.0.0.1:10251         0.0.0.0:*               LISTEN      23783/kube-schedule

$ curl -s http://127.0.0.1:10251/metrics |head
# HELP apiserver_audit_event_total Counter of audit events generated and sent to the audit backend.
# TYPE apiserver_audit_event_total counter
apiserver_audit_event_total 0
# HELP go_gc_duration_seconds A summary of the GC invocation durations.
# TYPE go_gc_duration_seconds summary
go_gc_duration_seconds{quantile="0"} 9.7715e-05
go_gc_duration_seconds{quantile="0.25"} 0.000107676
go_gc_duration_seconds{quantile="0.5"} 0.00017868
go_gc_duration_seconds{quantile="0.75"} 0.000262444
go_gc_duration_seconds{quantile="1"} 0.001205223

测试 kube-scheduler 集群的高可用

随便找一个或两个 master 节点，停掉 kube-scheduler 服务，看其它节点是否获取了 leader 权限（systemd 日志）。

查看当前的 leader

$ kubectl get endpoints kube-scheduler --namespace=kube-system  -o yaml
apiVersion: v1
kind: Endpoints
metadata:
  annotations:
    control-plane.alpha.kubernetes.io/leader: '{"holderIdentity":"m7-autocv-gpu01_7295c239-f2e9-11e8-8b5d-0cc47a2afc6a","leaseDurationSeconds":15,"acquireTime":"2018-11-28T08:41:50Z","renewTime":"2018-11-28T08:42:08Z","leaderTransitions":0}'
  creationTimestamp: 2018-11-28T08:41:50Z
  name: kube-scheduler
  namespace: kube-system
  resourceVersion: "1013"
  selfLink: /api/v1/namespaces/kube-system/endpoints/kube-scheduler
  uid: 73305545-f2e9-11e8-b65b-0cc47a2afc6a

可见，当前的 leader 为 m7-autocv-gpu01 节点。

06-4.scheduler集群