ceph集群搭建

Posted 2020-10-08

一、安装前准备：

1.系统：Centos7.4 x64

[[email protected] ~]# cat /etc/redhat-release
CentOS Linux release 7.4.1708 (Core)

2.主机：

主机名	地址	角色
ceph-node1	10.0.70.40	Deploy、mon1、osd*2
ceph-node2	10.0.70.41	mon1、osd*2
ceph-node3	10.0.70.42	mon1、osd*2

3.三台主机，每台主机有两个磁盘（磁盘大于100G）

[[email protected] ~]# lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
fd0           2:0    1     4K  0 disk
sda           8:0    0  1000G  0 disk
├─sda1        8:1    0     1G  0 part /boot
└─sda2        8:2    0   999G  0 part
├─cl-root 253:0    0    50G  0 lvm  /
├─cl-swap 253:1    0   7.9G  0 lvm  [SWAP]
└─cl-home 253:2    0 941.1G  0 lvm  /home
sdb           8:16   0  1000G  0 disk
sdc           8:32   0  1000G  0 disk

域名解析

ssh-keygen       #配置ceph-1到其他ceph节点的无密码访问
ssh-copy-id [email protected]
ssh-copy-id [email protected]
ssh-copy-id [email protected]

vim /etc/hosts    #拷贝解析到所有站点
10.0.70.40 ceph-node1
10.0.70.41 ceph-node2
10.0.70.42 ceph-node3
scp /etc/hosts [email protected]:/etc/
scp /etc/hosts [email protected]:/etc/

yum源配置

yum clean all
rm -rf /etc/yum.repos.d/*.repo
wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo
wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
sed -i ‘/aliyuncs/d‘ /etc/yum.repos.d/CentOS-Base.repo
sed -i ‘/aliyuncs/d‘ /etc/yum.repos.d/epel.repo
sed -i ‘s/$releasever/7/g‘ /etc/yum.repos.d/CentOS-Base.repo
vim /etc/yum.repos.d/ceph.repo
[ceph]
name=ceph
baseurl=http://mirrors.aliyun.com/ceph/rpm-jewel/el7/x86_64/
gpgcheck=0
[ceph-noarch]
name=cephnoarch
baseurl=http://mirrors.aliyun.com/ceph/rpm-jewel/el7/noarch/
gpgcheck=0
scp /etc/yum.repos.d/ceph.repo [email protected]:/etc/yum.repos.d/     #拷贝镜像到所有站点
scp /etc/yum.repos.d/ceph.repo [email protected]:/etc/yum.repos.d/
yum install hdparm  ceph ceph-radosgw rdate  -y    #安装ceph客户端(所有节点)
hdparm -W 0 /dev/sda      #关闭写缓存
rdate -s time-a.nist.gov        #同步时间
echo rdate -s time-a.nist.gov >> /etc/rc.d/rc.local
chmod +x /etc/rc.d/rc.local

二、集群部署

1.管理节点部署

yum install ceph-deploy -y
mkdir ~/cluster                     #创建my-cluster目录存放配置文件与秘钥
cd  ~/cluster                   #每次执行ceph-deploy命令最好都在这目录下执行
ceph-deploy new ceph-node1 ceph-node2 ceph-node3   #以node1为MON创建ceph.conf文件与秘钥
echo "osd_pool_default_size" = 2 >> ~/cluster/ceph.conf    #设置默认的副本数
echo public_network=10.0.70.20/24 >> ~/cluster/ceph.conf       #根据自己的IP配置向ceph.conf中添加public_network
echo mon_clock_drift_allowed = 2 >> ~/cluster/ceph.conf    #增大mon之间时差允许范围
cat  ~/cluster/ceph.conf
osd_pool_default_size = 2
public_network=10.0.70.20/24
mon_clock_drift_allowed = 2

开始部署monitor

ceph-deploy mon create-initial

开始部署OSD:

ceph-deploy --overwrite-conf osd prepare ceph-node1:/dev/sdb ceph-node1:/dev/sdc ceph-node2:/dev/sdb ceph-node2:/dev/sdc  ceph-node3:/dev/sdb ceph-node3:/dev/sdc   --zap-disk
ceph-deploy --overwrite-conf osd activate ceph-node1:/dev/sdb1 ceph-node1:/dev/sdc1  ceph-node2:/dev/sdb1 ceph-node2:/dev/sdc1  ceph-node3:/dev/sdb1 ceph-node3:/dev/sdc1

查看集群状态

[[email protected] cluster]# ceph -s
cluster 466e0a3e-f351-46f3-94a2-5ea976c26fd8
health HEALTH_WARN
15 pgs peering
2 pgs stuck unclean
too few PGs per OSD (21 < min 30)
monmap e1: 3 mons at {ceph-node1=10.0.70.40:6789/0,ceph-node2=10.0.70.41:6789/0,ceph-node3=10.0.70.42:6789/0}
election epoch 4, quorum 0,1,2 ceph-node1,ceph-node2,ceph-node3
osdmap e47: 6 osds: 6 up, 6 in; 9 remapped pgs
flags sortbitwise,require_jewel_osds
pgmap v125: 64 pgs, 1 pools, 0 bytes data, 0 objects
203 MB used, 5966 GB / 5967 GB avail
49 active+clean
9 remapped+peering
6 peering

查看OSD

[[email protected] cluster]# ceph osd tree
ID WEIGHT  TYPE NAME           UP/DOWN REWEIGHT PRIMARY-AFFINITY
-1 5.82715 root default
-2 1.94238     host ceph-node1
0 0.97119         osd.0            up  1.00000          1.00000
1 0.97119         osd.1            up  1.00000          1.00000
-3 1.94238     host ceph-node2
2 0.97119         osd.2            up  1.00000          1.00000
3 0.97119         osd.3            up  1.00000          1.00000
-4 1.94238     host ceph-node3
4 0.97119         osd.4            up  1.00000          1.00000
5 0.97119         osd.5            up  1.00000          1.00000

问题1：

[ceph-1][ERROR ] admin_socket: exception getting command descriptions: [Errno 2] No such file or directory

[ceph_deploy.mon][WARNIN] mon.ceph-1 monitor is not yet in quorum, tries left: 5

[ceph_deploy.mon][WARNIN] waiting 5 seconds before retrying

解决：

主机名和hosts对应即可

本文出自 “卫斯理” 博客，请务必保留此出处http://szk5043.blog.51cto.com/8456440/1969469