说明

最近由于需要部署有状态服务，没有云环境的支持，不能很好的通过cloud provider调用云磁盘EBS，所以在node节点部署了一套glusterFS分布式文件存储系统，用来提供非结构化数据存储,存储图片，大文件等等.

Kubernetes中使用GlusterFS作为持久化存储，要提供storageClass使用需要依赖Heketi工具。Heketi是一个具有resetful接口的glusterfs管理程序，作为kubernetes的Storage存储的external provisioner。 “Heketi提供了一个RESTful管理界面，可用于管理GlusterFS卷的生命周期。借助Heketi，像OpenStack Manila，Kubernetes和OpenShift这样的云服务可以动态地配置GlusterFS卷和任何支持的持久性类型。Heketi将自动确定整个集群的brick位置，确保将brick及其副本放置在不同的故障域中。Heketi还支持任意数量的GlusterFS集群，允许云服务提供网络文件存储，而不受限于单个GlusterFS集群。

heketi：提供基于RESTful接口管理glusterfs的功能，可以方便的创建集群管理glusterfs的node，device，volume；与k8s结合可以创建动态的PV，扩展glusterfs存储的动态管理功能。主要用来管理glusterFS volume的生命周期，初始化时候就要分配好裸磁盘(未格式化)设备.

注意事项

• 安装Glusterfs客户端：每个kubernetes集群的节点需要安装gulsterfs的客户端，如glusterfs-cli,glusterfs-fuse.主要用于在每个node节点挂载volume。
• 加载内核模块：每个kubernetes集群的节点运行modprobe dm_thin_pool，加载内核模块。
• 高可用(至少三个节点)：至少需要节点用来部署glusterfs集群，并且这3个节点每个节点需要至少一个空余的磁盘。

基础设施要求:

• 正在运行的Kubernetes集群，至少有三个node节点，每个节点至少有一个可用的裸块设备（如EBS卷或本地磁盘,就是没有格式化的）.
• 用于运行GlusterFS节点必须为GlusterFS通信打开相应的端口（如果开启了防火墙的情况下，没开防火墙就不需要这些操作）。在每个节点上运行以下命令

$ sudo iptables -N heketi
$ sudo iptables -A heketi -p tcp -m state --state NEW -m tcp --dport 24007 -j $ sudo ACCEPT
$ sudo iptables -A heketi -p tcp -m state --state NEW -m tcp --dport 24008 -j ACCEPT
$ sudo iptables -A heketi -p tcp -m state --state NEW -m tcp --dport 2222 -j ACCEPT
$ sudo iptables -A heketi -p tcp -m state --state NEW -m multiport --dports 49152:49251 -j ACCEPT
$ sudo service iptables save

Heketi安装

Heketi是由golang编写，直接静态编译运行二进制即可也可以通过yum安装以及docker部署，主要会产生db文件存储cluster、node、device、volume等信息。

#/bin/bash
wget -c https://github.com/heketi/heketi/releases/download/v8.0.0/heketi-v8.0.0.linux.amd64.tar.gz
tar xzf heketi-v8.0.0.linux.amd64.tar.gz
mkdir -pv /data/heketi/{bin,conf,data}
cp heketi/heketi.json /data/heketi/conf/
cp heketi/{heketi,heketi-cli} /data/heketi/bin/
mv heketi-v8.0.0.linux.amd64.tar.gz /tmp/

创建ssh-key

我们glusterFS部署在k8s集群外，所以heketi通过ssh管理glusterFS。需要创建免秘钥登陆到所有glusterFS节点。

$ sudo  ssh-keygen -f /data/heketi/conf/heketi_key -t rsa -N ''
# 将公钥放到所有GlusterFS节点
$ sudo ssh-copy-id -i /data/heketi/conf/heketi_key.pub root@node1
$ ...

配置文件修改

参考https://github.com/heketi/heketi/blob/master/client/cli/go/topology-sample.json

{
  "_port_comment": "Heketi Server Port Number",
  "port": "18080",

  "_use_auth": "Enable JWT authorization. Please enable for deployment",
  # 开启认证
  "use_auth": true,

  "_jwt": "Private keys for access",
  "jwt": {
    "_admin": "Admin has access to all APIs",
    # admin用户的key
    "admin": {
      "key": "adminkey"
    },
    "_user": "User only has access to /volumes endpoint",
    "user": {
      "key": "userkey"
    }
  },

  "_glusterfs_comment": "GlusterFS Configuration",
  "glusterfs": {
    "_executor_comment": [
      "Execute plugin. Possible choices: mock, ssh",
      "mock: This setting is used for testing and development.",
      "      It will not send commands to any node.",
      "ssh:  This setting will notify Heketi to ssh to the nodes.",
      "      It will need the values in sshexec to be configured.",
      "kubernetes: Communicate with GlusterFS containers over",
      "            Kubernetes exec api."
    ],
    # 修改执行插件为ssh，并配置ssh的所需证书，注意要能对集群中的机器免密ssh登陆，使用ssh-copy-id把pub key拷到每台glusterfs服务器上
    # heketi 通过ssh远程控制glusterfs的volume周期管理。
    "executor": "ssh",

    "_sshexec_comment": "SSH username and private key file information",
    "sshexec": {
       # ssh-keygen 生成，然后拷贝到glusterfs所在机器
      "keyfile": "/data/heketi/conf/heketi_key",
      "user": "root",
      "port": "22",
      # 每创建一个volume，heketi会刷新fstab配置保证机器自启动会自动挂载。
      "fstab": "/etc/fstab"
    },

    "_kubeexec_comment": "Kubernetes configuration",
    "kubeexec": {
      "host" :"https://kubernetes.host:8443",
      "cert" : "/path/to/crt.file",
      "insecure": false,
      "user": "kubernetes username",
      "password": "password for kubernetes user",
      "namespace": "OpenShift project or Kubernetes namespace",
      "fstab": "Optional: Specify fstab file on node.  Default is /etc/fstab"
    },

    "_db_comment": "Database file name",
    # 修改heketi默认db路径
    "db": "/data/heketi/data/heketi.db",

    "_loglevel_comment": [
      "Set log level. Choices are:",
      "  none, critical, error, warning, info, debug",
      "Default is warning"
    ],
    "loglevel" : "debug"
  }
}

注意：

• 需要说明的是，heketi有三种executor，分别为mock、ssh、kubernetes，建议在测试环境使用mock，生产环境使用ssh，当glusterfs以容器的方式部署在kubernetes上时，才使用kubernetes。我们这里将glusterfs和heketi独立部署，使用ssh的方式。
• 使用docker部署的时候，还需将/var/lib/heketi/mounts 挂载至容器里面, heketi 会将此目录作为 gluster volume的挂载点。

[root@k8s1 ~]# more /etc/fstab
#
# /etc/fstab
# Created by anaconda on Sun Oct 15 15:19:00 2017
#
# Accessible filesystems, by reference, are maintained under '/dev/disk'
# See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info
#
UUID=eb448abb-3012-4d8d-bcde-94434d586a31 /                       ext4    defaults        1 1
UUID=63b64b3a-2961-4798-b7f1-cc97484ee49f /data                   ext4    defaults        1 1
/dev/mapper/vg_fd3a11426117508af77aa38e9565ce65-brick_4b7854340fa8ee78106b1b1446abf078 /var/lib/heketi/mounts/vg_fd3a11426117508af77aa38e9565ce65/brick_4b
7854340fa8ee78106b1b1446abf078 xfs rw,inode64,noatime,nouuid 1 2
/dev/mapper/vg_fd3a11426117508af77aa38e9565ce65-brick_1904166cdf0846ee649ccefe66ce1e50 /var/lib/heketi/mounts/vg_fd3a11426117508af77aa38e9565ce65/brick_19
04166cdf0846ee649ccefe66ce1e50 xfs rw,inode64,noatime,nouuid 1 2

systemd配置

# cat /usr/lib/system/systemd/heketi.service
[Unit]
Description=RESTful based volume management framework for GlusterFS
Before=network-online.target
After=network-online.target
Documentation=https://github.com/heketi/heketi

[Service]
Type=simple
LimitNOFILE=65536
ExecStart=/data/heketi/bin/heketi --config=/data/heketi/conf/heketi.json
KillMode=process
Restart=on-failure
RestartSec=5
SuccessExitStatus=15
StandardOutput=syslog
StandardError=syslog

[Install]
WantedBy=multi-user.target

启动heketi服务

$ sudo systemctl start heketi
$sudo systemctl enable heketi
$sudo systemctl status heketi

Heketi管理GlusterFS

添加cluster

$ sudo heketi-cli --user admin --server http://10.111.209.188:18080 --secret adminkey --json  cluster create
{"id":"d102a74079dd79aceb3c70d6a7e8b7c4","nodes":[],"volumes":[],"block":true,"file":true,"blockvolumes":[]}

将3个glusterfs节点作为node添加到cluster

由于我们开启了heketi认证，所以每次执行heketi-cli操作时，都需要带上一堆的认证字段，比较麻烦，我在这里创建一个别名来避免相关操作：

$ sudo  alias heketi-cli='heketi-cli --server "[http://192.168.75.175:18080"](http://192.168.75.175:18080/) --user "admin" --secret "adminkey"'

添加节点

$sudo heketi-cli --json node add --cluster "d102a74079dd79aceb3c70d6a7e8b7c4" --management-host-name k8s1 --storage-host-name 10.111.209.188  --zone 1

$sudo heketi-cli --json node add --cluster "d102a74079dd79aceb3c70d6a7e8b7c4" --management-host-name k8s1 --storage-host-name 10.111.209.189 --zone 1

$sudo heketi-cli --json node add --cluster "d102a74079dd79aceb3c70d6a7e8b7c4" --management-host-name k8s3 --storage-host-name  10.111.209.190 --zone 1

注意: 如果heketi 运行账号为非root，需要保证有sudo权限。

添加device

机器只是作为gluster的运行单元，volume是基于device创建的。同时需要特别说明的是，目前heketi仅支持使用裸分区或裸磁盘(未格式化)添加为device，不支持文件系统。

# --node参数给出的id是上一步创建node时生成的，这里只给出一个添加的示例，实际配置中，要添加每一个节点的每一块用于存储的硬盘`

$ sudo  heketi-cli  -json device add -name="/dev/vdc1" --node ``"2e4dc73fb2657586e7a9e64e39c8f01a"

$ sudo heketi-cli node list`

Id:2e4dc73fb2657586e7a9e64e39c8f01a     Cluster:d102a74079dd79aceb3c70d6a7e8b7c4

Id:43a661a917c42dfe9b6659b8c9d848e9     Cluster:d102a74079dd79aceb3c70d6a7e8b7c4

Id:ce5a9e2983d067051b6ccad0a4bd0988     Cluster:d102a74079dd79aceb3c70d6a7e8b7c4

生产实际配置

以上ad-hoc命令均可通过配置文件创建然后导入:

$ sudo  cat /data/heketi/conf/topology.json

{
    "clusters": [
        {
            "nodes": [
                {
                    "node": {
                        "hostnames": {
                            "manage": [
                                "k8s1"
                            ],
                            "storage": [
                                "10.111.209.188"
                            ]
                        },
                        "zone": 1
                    },
                    "devices": [
                        "/dev/vdc1"
                    ]
                },
                {
                    "node": {
                        "hostnames": {
                            "manage": [
                                "k8s2"
                            ],
                            "storage": [
                                "10.111.209.189"
                            ]
                        },
                        "zone": 1
                    },
                    "devices": [
                        "/dev/vdc1"
                    ]
                },
                {
                    "node": {
                        "hostnames": {
                            "manage": [
                                "k8s3"
                            ],
                            "storage": [
                                "10.111.209.190"
                            ]
                        },
                        "zone": 1
                    },
                    "devices": [
                        "/dev/vdc1"
                    ]
                }             
            ]
        }
    ]
}

创建:

$ sudo heketi-cli  topology load --json topology.json

添加volume

这里仅仅是做一个测试，实际使用中，会由kubernetes自动创建pv.

创建一个大小为3G，副本为2的volume

$ sudo  heketi-cli volume create --size 3 --replica 2

Name: vol_685f9aea1896f53f30a22a9d15de1680

Size: 3

Volume Id: 685f9aea1896f53f30a22a9d15de1680

Cluster Id: d102a74079dd79aceb3c70d6a7e8b7c4

Mount: 10.111.209.188:vol_685f9aea1896f53f30a22a9d15de1680

Mount Options: backup-volfile-servers=10.111.209.190,10.111.209.189

Block: false

Free Size: 0

Reserved Size: 0

Block Hosting Restriction: (none)

Block Volumes: []

Durability Type: replicate

Distributed+Replica: 2

kubernetes storageclass 配置

创建storageclass

添加storageclass-glusterfs.yaml文件，内容如下：

---
apiVersion: v1
kind: Secret
metadata:
  name: heketi-secret
  namespace: default
data:
  # base64 encoded password. E.g.: echo -n "mypassword" | base64
  key: YWRtaW5rZXk=
type: kubernetes.io/glusterfs

---
apiVersion: storage.k8s.io/v1beta1
kind: StorageClass
metadata:
  name: glusterfs
provisioner: kubernetes.io/glusterfs
allowVolumeExpansion: true
parameters:
  resturl: "http://10.111.209.188:18080"
  clusterid: "6fd6bf78b84315e12abcf8b6db6b1a40"
  restauthenabled: "true"
  restuser: "admin"
  #secretNamespace: "default"
  #secretName: "heketi-secret"
  restuserkey: "adminkey"
  gidMin: "40000"
  gidMax: "50000"
  volumetype: "replicate:2"
# kubectl apply -f glusterfs-storageclass.yml
secret/heketi-secret created
storageclass.storage.k8s.io/glusterfs created
# kubectl get sc
NAME                  PROVISIONER               AGE
glusterfs (default)   kubernetes.io/glusterfs   2m19s

注意:

• storageclass.beta.kubernetes.io/is-default-class: "true" #表示此storageClass作为default sc，创建pvc不指定sc时，默认使用此sc.
• reclaimPolicy: Retain #表示pv回收策略为保留，删除pvc时将不删除pv。

更详细的用法参考：https://kubernetes.io/docs/concepts/storage/storage-classes/#glusterfs

创建pvc

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: glusterfs-mysql1
  namespace: default
  annotations:
    volume.beta.kubernetes.io/storage-class: "glusterfs"
spec:
  accessModes:
  - ReadWriteMany
  resources:
    requests:
      storage: 1Gi
      
# kubectl create -f glusterfs-pvc.yaml
persistentvolumeclaim/glusterfs-mysql1 created
# kubectl get pv,pvc
NAME                                                        CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                      STORAGECLASS   REASON   AGE
persistentvolume/pvc-11aa48c5-fe0f-11e8-9803-00163e13b711   1Gi        RWX            Retain           Bound    default/glusterfs-mysql1   glusterfs               2m

NAME                                     STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE
persistentvolumeclaim/glusterfs-mysql1   Bound    pvc-11aa48c5-fe0f-11e8-9803-00163e13b711   1Gi        RWX            glusterfs      2m20s

创建pod，使用pvc

---
kind: Deployment
apiVersion: extensions/v1beta1
metadata:
  name: mysql
  namespace: default
spec:
  replicas: 1
  template:
    metadata:
      labels:
        name: mysql
    spec:
      containers:
      - name: mysql
        image: mysql:5.7
        imagePullPolicy: IfNotPresent
        env:
        - name: MYSQL_ROOT_PASSWORD
          value: root123456
        ports:
          - containerPort: 3306
        volumeMounts:
        - name: gluster-mysql-data
          mountPath: "/var/lib/mysql"
      volumes:
        - name: gluster-mysql-data
          persistentVolumeClaim:
            claimName: glusterfs-mysql1
            
# kubectl create -f /etc/kubernetes/mysql-deployment.yaml
deployment.extensions/mysql created
# kubectl exec -it mysql-84786cf494-hb2ss -- df -PHT /var/lib/mysql
Filesystem                                          Type            Size  Used Avail Use% Mounted on
10.111.209.188:vol_426e62366141c789dac33f2e68dfb13b fuse.glusterfs  1.1G  266M  798M  25% /var/lib/mysql

创建statefulset

---
apiVersion: v1
kind: Service
metadata:
  name: nginx
  labels:
    app: nginx
spec:
  ports:
  - port: 80
    name: web
  clusterIP: None
  selector:
    app: nginx
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: nginx
spec:
  selector:
    matchLabels:
      app: nginx # has to match .spec.template.metadata.labels
  serviceName: "nginx"
  replicas: 3 # by default is 1
  template:
    metadata:
      labels:
        app: nginx # has to match .spec.selector.matchLabels
    spec:
      terminationGracePeriodSeconds: 10
      containers:
      - name: nginx
        image: nginx
        ports:
        - containerPort: 80
          name: web
        volumeMounts:
        - name: www
          mountPath: /usr/share/nginx/html
  volumeClaimTemplates:
  - metadata:
      name: www
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "glusterfs"
      resources:
        requests:
          storage: 1Gi
# kubectl apply -f nginx-statefulset.yml
service/nginx created
statefulset.apps/nginx created


# kubectl get pod,pv,pvc
NAME           READY   STATUS    RESTARTS   AGE
pod/nginx-0    1/1     Running   0          116s
pod/nginx-1    1/1     Running   0          98s
pod/nginx-2    1/1     Running   0          91s

NAME                                                        CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                 STORAGECLASS   REASON   AGE
persistentvolume/pvc-5ac3eba9-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            Retain           Bound    default/www-nginx-0   glusterfs               99s
persistentvolume/pvc-65f27519-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            Retain           Bound    default/www-nginx-1   glusterfs               93s
persistentvolume/pvc-69b31512-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            Retain           Bound    default/www-nginx-2   glusterfs               86s

NAME                                STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE
persistentvolumeclaim/www-nginx-0   Bound    pvc-5ac3eba9-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            glusterfs      116s
persistentvolumeclaim/www-nginx-1   Bound    pvc-65f27519-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            glusterfs      98s
persistentvolumeclaim/www-nginx-2   Bound    pvc-69b31512-fe12-11e8-a19e-00163e14b18f   1Gi        RWO            glusterfs      91s

我们可以看到RECLAIM POLICY: Retain ,经过测试

• 删除pvc，pv status会变成Released状态，且不会被删除
• 删除pv, 通过heketi-cli volume list查看，volume不会被删除

kubernetes pv和gluster volume不一致时,可使用heketi来统一管理volume.此文档heketi和glusterfs都在kubernetes集群外部署。对于支持AWS EBS的磁盘，可通过EBS storageClass方式将glusterFS heketi部署在容器中管理.参考https://github.com/gluster/gluster-kubernetes

参考文档

• https://kubernetes.io/docs/concepts/workloads/controllers/statefulset/
• https://kubernetes.io/docs/concepts/storage/storage-classes/#glusterfs
• https://github.com/heketi/heketi/blob/master/docs/admin/readme.md
• https://www.cnblogs.com/breezey/p/8849466.html