centos7搭建ceph集群

一、服务器规划

主机名	主机IP	磁盘配比	角色
node1	public-ip：10.0.0.130 cluster-ip:192.168.2.130	sda,sdb,sdc sda是系统盘，另外两块数据盘	ceph-deploy,monitor,mgr,osd
node2	public-ip：10.0.0.131 cluster-ip:192.168.2.131	sda,sdb,sdc sda是系统盘，另外两块数据盘	monitor,mgr,osd
node3	public-ip：10.0.0.132 cluster-ip:192.168.2.132	sda,sdb,sdc sda是系统盘，另外两块数据盘	monitor,mgr,osd

二、设置主机名

主机名设置，三台主机分别执行属于自己的命令

node1

[root@localhost ~]# hostnamectl set-hostname node1
[root@localhost ~]# hostname node1

node2

[root@localhost ~]# hostnamectl set-hostname node2
[root@localhost ~]# hostname node2

node3

[root@localhost ~]# hostnamectl set-hostname node3
[root@localhost ~]# hostname node3

执行完毕后要想看到效果，需要关闭当前命令行窗口，重新打开即可看到设置效果

三、设置hosts文件

在3台机器上都执行下面命令，添加映射

echo "10.0.0.130 node1 " >> /etc/hosts
echo "10.0.0.131 node2 " >> /etc/hosts
echo "10.0.0.132 node3 " >> /etc/hosts

四、创建用户并设置免密登录

创建用户（三台机器上都运行）

useradd -d /home/admin -m admin
echo "" | passwd admin --stdin
#sudo权限
echo "admin ALL = (root) NOPASSWD:ALL" | sudo tee /etc/sudoers.d/admin
sudo chmod  /etc/sudoers.d/admin

设置免密登录  （只在node1上执行）

[root@node1 ~]# su - admin
[admin@node1 ~]$ ssh-keygen
Generating public/private rsa key pair.
Enter file in which to save the key (/home/admin/.ssh/id_rsa):
Created directory '/home/admin/.ssh'.
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in /home/admin/.ssh/id_rsa.
Your public key has been saved in /home/admin/.ssh/id_rsa.pub.
The key fingerprint is:
SHA256:qfWhuboKeoHQOOMLOIB5tjK1RPjgw/Csl4r6A1FiJYA admin@admin.ops5.bbdops.com
The key's randomart image is:
+---[RSA 2048]----+
|+o..             |
|E.+              |
|*%               |
|X+X      .       |
|=@.+    S .      |
|X.*    o + .     |
|oBo.  . o .      |
|ooo.     .       |
|+o....oo.        |
+----[SHA256]-----+
[admin@node1 ~]$ ssh-copy-id admin@node1
[admin@node1 ~]$ ssh-copy-id admin@node2
[admin@node1 ~]$ ssh-copy-id admin@node3

五、配置时间同步

三台都执行

yum -y install ntpdate
ntpdate -u  cn.ntp.org.cn

crontab -e
*/ * * * * ntpdate -u  cn.ntp.org.cn > /dev/null >&

systemctl reload crond.service

六、安装ceph-deploy并安装ceph软件包

配置ceph清华源

cat > /etc/yum.repos.d/ceph.repo<<'EOF'
[Ceph]
name=Ceph packages for $basearch
baseurl=https://mirror.tuna.tsinghua.edu.cn/ceph/rpm-mimic/el7/$basearch
enabled=
gpgcheck=
type=rpm-md
gpgkey=https://mirror.tuna.tsinghua.edu.cn/ceph/keys/release.asc
priority=
[Ceph-noarch]
name=Ceph noarch packages
baseurl=https://mirror.tuna.tsinghua.edu.cn/ceph/rpm-mimic/el7/noarch
enabled=
gpgcheck=
type=rpm-md
gpgkey=https://mirror.tuna.tsinghua.edu.cn/ceph/keys/release.asc
priority=
[ceph-source]
name=Ceph source packages
baseurl=https://mirror.tuna.tsinghua.edu.cn/ceph/rpm-mimic/el7/SRPMS
enabled=
gpgcheck=
type=rpm-md
gpgkey=https://mirror.tuna.tsinghua.edu.cn/ceph/keys/release.asc
priority=
EOF

安装ceph-deploy

[root@node1 ~]# sudo yum install ceph-deploy

初始化mon点

ceph需要epel源的包，所以安装的节点都需要yum install epel-release

[admin@node1 ~]$ mkdir my-cluster
[admin@node1 ~]$ cd my-cluster
# new
[admin@node1 my-cluster]$ ceph-deploy new node1 node2 node3
Traceback (most recent call last):
  File "/bin/ceph-deploy", line , in <module>
    from ceph_deploy.cli import main
  File "/usr/lib/python2.7/site-packages/ceph_deploy/cli.py", line , in <module>
    import pkg_resources
ImportError: No module named pkg_resources
#以上出现报错，是因为没有pip，安装pip
[admin@node1 my-cluster]$ sudo yum install epel-release
[admin@node1 my-cluster]$ sudo yum install python-pip
#重新初始化
[admin@node1 my-cluster]$ ceph-deploy new node1 node2 node3
[admin@node1 my-cluster]$ ls
ceph.conf  ceph-deploy-ceph.log  ceph.mon.keyring
[admin@node1 my-cluster]$ cat ceph.conf
[global]
fsid = a1132f78-cdc5-43d0-9ead-5b590c60c53d
mon_initial_members = node1, node2, node3
mon_host = 10.28.103.211,10.28.103.212,10.28.103.213
auth_cluster_required = cephx
auth_service_required = cephx
auth_client_required = cephx

修改ceph.conf，添加如下配置

public network = 10.28.103.0/
cluster network = 172.30.103.0/
osd pool default size       =
osd pool default min size   =
osd pool default pg num     =
osd pool default pgp num    =
osd pool default crush rule =
osd crush chooseleaf type   =
max open files              =
ms bind ipv6                = false
[mon]
mon clock drift allowed      =
mon clock drift warn backoff =
mon osd full ratio           = .
mon osd nearfull ratio       = .
mon osd down out interval    =
mon osd report timeout       =
mon allow pool delete      = true
[osd]
osd recovery max active      =
osd max backfills            =
osd max scrubs               =
osd mkfs type = xfs
osd mkfs options xfs = -f -i size=
osd mount options xfs = rw,noatime,inode64,logbsize=256k,delaylog
filestore max sync interval  =
osd op threads               = 

安装Ceph软件到指定节点

[admin@node1 my-cluster]$ ceph-deploy install --no-adjust-repos node1 node2 node3

--no-adjust-repos是直接使用本地源，不生成官方源。

部署初始的monitors，并获得keys

[admin@node1 my-cluster]$ ceph-deploy mon create-initial

做完这一步，在当前目录下就会看到有如下的keyrings：

[admin@node1 my-cluster]$ ls
ceph.bootstrap-mds.keyring  ceph.bootstrap-osd.keyring  ceph.client.admin.keyring  ceph-deploy-ceph.log
ceph.bootstrap-mgr.keyring  ceph.bootstrap-rgw.keyring  ceph.conf                  ceph.mon.keyring

将配置文件和密钥复制到集群各节点

配置文件就是生成的ceph.conf，而密钥是ceph.client.admin.keyring，当使用ceph客户端连接至ceph集群时需要使用的密默认密钥，这里我们所有节点都要复制，命令如下。

[admin@node1 my-cluster]$ ceph-deploy admin node1 node2 node3

七、部署ceph-mgr

#在L版本的`Ceph`中新增了`manager daemon`，如下命令部署一个`Manager`守护进程
[admin@node1 my-cluster]$ ceph-deploy mgr create node1 

八、创建osd

在node1上执行以下命令

#用法：ceph-deploy osd create –data {device} {ceph-node}
ceph-deploy osd create --data /dev/sdb node1
ceph-deploy osd create --data /dev/sdb node2
ceph-deploy osd create --data /dev/sdb node3
ceph-deploy osd create --data /dev/sdc node1
ceph-deploy osd create --data /dev/sdc node2
ceph-deploy osd create --data /dev/sdc node3

如果报错，记得用root执行

检查osd状态

[admin@node1 ~]$ sudo ceph health
HEALTH_OK
[admin@node1 ~]$ sudo ceph -s
  cluster:
    id:     af6bf549-45be-419c-92a4-8797c9a36ee8
    health: HEALTH_OK

  services:
    mon:  daemons, quorum node1,node2,node3
    mgr: node1(active)
    osd:  osds:  up,  in

  data:
    pools:    pools,  pgs
    objects:   objects,  B
    usage:   6.0 GiB used,  GiB /  GiB avail
    pgs:
 

默认情况下ceph.client.admin.keyring文件的权限为600，属主和属组为root，如果在集群内节点使用cephadmin用户直接直接ceph命令，将会提示无法找到/etc/ceph/ceph.client.admin.keyring文件，因为权限不足。

如果使用sudo ceph不存在此问题，为方便直接使用ceph命令，可将权限设置为644。在集群节点上面node1 admin用户下执行下面命令。

[admin@node1 my-cluster]$ ceph -s
-- ::36.062 7f52d08e0700 - auth: unable to find a keyring on /etc/ceph/ceph.client.admin.keyring,/etc/ceph/ceph.keyring,/etc/ceph/keyring,/etc/ceph/keyring.bin,: () No such file or directory
-- ::36.062 7f52d08e0700 - monclient: ERROR: missing keyring, cannot use cephx for authentication
[errno ] error connecting to the cluster
[admin@node1 my-cluster]$ sudo chmod  /etc/ceph/ceph.client.admin.keyring
[admin@node1 my-cluster]$ ceph -s
  cluster:
    id:     af6bf549-45be-419c-92a4-8797c9a36ee8
    health: HEALTH_OK

  services:
    mon:  daemons, quorum node1,node2,node3
    mgr: node1(active)
    osd:  osds:  up,  in

  data:
    pools:    pools,  pgs
    objects:   objects,  B
    usage:   6.1 GiB used,  GiB /  GiB avail
    pgs:     

[admin@node1 my-cluster]$ 

查看osds

[admin@node1 ~]$ sudo ceph osd tree
ID CLASS WEIGHT  TYPE NAME      STATUS REWEIGHT PRI-AFF
-       0.11151 root default
-       0.03717     host node1
    hdd 0.01859         osd.      up  1.00000 1.00000
    hdd 0.01859         osd.      up  1.00000 1.00000
-       0.03717     host node2
    hdd 0.01859         osd.      up  1.00000 1.00000
    hdd 0.01859         osd.      up  1.00000 1.00000
-       0.03717     host node3
    hdd 0.01859         osd.      up  1.00000 1.00000
    hdd 0.01859         osd.      up  1.00000 1.00000 

九、开启MGR监控模块

方式一：命令操作

ceph mgr module enable dashboard

如果以上操作报错如下：

Error ENOENT: all mgr daemons do not support module 'dashboard', pass --force to force enablement

则因为没有安装ceph-mgr-dashboard，在mgr的节点上安装。

yum install ceph-mgr-dashboard

方式二：配置文件

# 编辑ceph.conf文件
vi ceph.conf
[mon]
mgr initial modules = dashboard
#推送配置
[admin@admin my-cluster]$ ceph-deploy --overwrite-conf config push node1 node2 node3
#重启mgr
 sudo systemctl restart ceph-mgr@node1

web登录配置

默认情况下，仪表板的所有HTTP连接均使用SSL/TLS进行保护。

#要快速启动并运行仪表板，可以使用以下内置命令生成并安装自签名证书:
[root@node1 my-cluster]# ceph dashboard create-self-signed-cert
Self-signed certificate created

#创建具有管理员角色的用户:
[root@node1 my-cluster]# ceph dashboard set-login-credentials admin admin
Username and password updated
#查看ceph-mgr服务:
[root@node1 my-cluster]# ceph mgr services
{
    "dashboard": "https://node1:8443/"
}

以上配置完成后，浏览器输入https://node1:8443输入用户名admin，密码admin登录即可查看

参考链接：

https://www.sysit.cn/blog/post/sysit/Ceph%E5%AE%89%E8%A3%85%E9%85%8D%E7%BD%AE%E6%89%8B%E5%86%8C

https://boke.wsfnk.com/archives/1163.html

https://www.linux-note.cn/?p=85