CEPH快速部署(Centos7+Jewel)
ceph介绍Ceph是统一存储系统，支持三种接口。
Object：有原生的API，而且也兼容Swift和S3的API
Block：支持精简配置、快照、克隆
File：Posix接口，支持快照
Ceph也是分布式存储系统，它的特点是：
高扩展性：使用普通x86服务器，支持10~1000台服务器，支持TB到PB级的扩展。
高可靠性：没有单点故障，多数据副本，自动管理，自动修复。
高性能：数据分布均衡，并行化度高。对于objects storage和block storage,不需要元数据服务器。
ceph架构
组件
http://s3.运维网.com/wyfs02/M01/8B/3C/wKioL1hHwnTQYV9NAAJmDpAPLSs348.png
Ceph的底层是RADOS，它的意思是“A reliable, autonomous, distributed object storage”。 RADOS由两个组件组成：
OSD： Object Storage Device，提供存储资源。
Monitor：维护整个Ceph集群的全局状态。
环境：三台装有Centos7的 主机，每个主机有三个磁盘（虚拟机磁盘要大于100G）
# cat /etc/redhat-release
CentOS Linux release 7.2.1511 (Core)
# lsblk
NAME MAJ:MIN RMSIZE RO TYPE MOUNTPOINT
fd0 2:0 1 4K0 disk
sda 8:0 0 20G0 disk
├─sda1 8:1 0500M0 part /boot
└─sda2 8:2 0 19.5G0 part
├─centos-root 253:0 0 18.5G0 lvm/
└─centos-swap 253:1 0 1G0 lvm
sdb 8:16 0200G0 disk
sdc 8:32 0200G0 disk
sdd 8:48 0200G0 disk
sr0 11:0 1603M0 rom
修改主机名并添加hosts
vim /etc/hostname #由于7和6.5修改主机名的方法不一样，这里举个例子
mon-1
# cat /etc/hosts
192.168.50.123mon-1
192.168.50.124osd-1
192.168.50.125osd-2 配置ssh无密码登陆
# ssh-keygen-t rsa -P ''
Generating public/private rsa key pair.
Enter file in which to save the key (/root/.ssh/id_rsa):
Your identification has been saved in /root/.ssh/id_rsa.
Your public key has been saved in /root/.ssh/id_rsa.pub.
The key fingerprint is:
62:b0:4c:aa:e5:37:92:89:4d:db:c3:38:e2:f1:2a:d6 root@admin-node
The key's randomart image is:
+--[ RSA 2048]----+
| |
| |
| o |
| + o |
|+ o o S |
| B B . . |
|+.@ * |
|oooE o |
|oo.. |
+-----------------+
ssh-copy-id mon-1
ssh-copy-id osd-1
ssh-copy-id osd-2
2.集群配置
主机IP功能
http://s2.运维网.com/wyfs02/M01/8B/40/wKiom1hH40LQm4yrAAAv7KPvlzY871.png
3.环境清理
如果之前部署失败了，不必删除ceph客户端，或者重新搭建虚拟机，只需要在每个节点上执行如下指令即可将环境清理至刚安装完ceph客户端时的状态！强烈建议在旧集群上搭建之前清理干净环境，否则会发生各种异常情况。
ps aux|grep ceph |awk '{print $2}'|xargs kill -9
ps -ef|grep ceph
#确保此时所有ceph进程都已经关闭！！！如果没有关闭，多执行几次。
umount /var/lib/ceph/osd/*
rm -rf /var/lib/ceph/osd/*
rm -rf /var/lib/ceph/mon/*
rm -rf /var/lib/ceph/mds/*
rm -rf /var/lib/ceph/bootstrap-mds/*
rm -rf /var/lib/ceph/bootstrap-osd/*
rm -rf /var/lib/ceph/bootstrap-mon/*
rm -rf /var/lib/ceph/tmp/*
rm -rf /etc/ceph/*
rm -rf /var/run/ceph/*
安装部署流程
yum源及ceph的安装
需要在每个主机上执行以下命令
yum clean all
rm -rf /etc/yum.repos.d/*.repo
wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo
wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
sed -i '/aliyuncs/d' /etc/yum.repos.d/CentOS-Base.repo
sed -i '/aliyuncs/d' /etc/yum.repos.d/epel.repo
sed -i 's/$releasever/7.2.1511/g' /etc/yum.repos.d/CentOS-Base.repo 增加ceph源
vim /etc/yum.repos.d/ceph.repo
name=ceph
baseurl=http://mirrors.aliyun.com/ceph/rpm-jewel/el7/x86_64/
gpgcheck=0
name=cephnoarch
baseurl=http://mirrors.aliyun.com/ceph/rpm-jewel/el7/noarch/
gpgcheck=0 安装ceph客户端
yum -y install ceph ceph-radosgw ntp ntpdate
关闭selinux和firewalld
sed -i 's/SELINUX=.*/SELINUX=disabled/' /etc/selinux/config #重启机器
systemctl stop firewalld
systemctl disable firewalld 同步各个节点时间：
ntpdate-s time-a.nist.gov
echo ntpdate-s time-a.nist.gov >> /etc/rc.d/rc.local
echo "01 01 * * * /usr/sbin/ntpdate -s time-a.nist.gov >> /dev/null 2>&1" >> /etc/crontab 开始部署
在部署节点(mon-1)安装ceph-deploy，下文的部署节点统一指mon-1:
# yum -y install ceph-deploy
# ceph-deploy --version
1.5.36
# ceph -v
ceph version 10.2.3 (ecc23778eb545d8dd55e2e4735b53cc93f92e65b) 在部署节点创建部署目录
# mkdir cluster
# cd cluster
# ceph-deploy new mon-1 osd-1 osd-2
found configuration file at: /root/.cephdeploy.conf
Invoked (1.5.36): /usr/bin/ceph-deploy new mon-1 osd-1 osd-2
ceph-deploy options:
username : None
func :
verbose : False
overwrite_conf : False
quiet : False
cd_conf :
cluster : ceph
ssh_copykey : True
mon : ['mon-1', 'osd-1', 'osd-2']
public_network : None
ceph_conf : None
cluster_network : None
default_release : False
fsid : None
Creating new cluster named ceph
making sure passwordless SSH succeeds
Writing initial config to ceph.conf... #如果能准确无误的走到这一步就算成功了 目录内容如下
# ll
total 12
-rw-r--r-- 1 root root197 Dec6 02:24 ceph.conf
-rw-r--r-- 1 root root 2921 Dec6 02:24 ceph-deploy-ceph.log
-rw------- 1 root root 73 Dec6 02:24 ceph.mon.keyring 根据自己的IP配置向ceph.conf中添加public_network，并稍微增大mon之间时差允许范围(默认为0.05s，现改为2s)：
# echo public_network=192.168.50.0/24 >>ceph.conf
# echo mon_clock_drift_allowed = 2 >> ceph.conf
# cat ceph.conf
fsid = 0865fe85-1655-4208-bed6-274cae945746
mon_initial_members = mon-1, osd-1, osd-2
mon_host = 192.168.50.123,192.168.50.124,192.168.50.125
auth_cluster_required = cephx
auth_service_required = cephx
auth_client_required = cephx
public_network=192.168.50.0/24
mon_clock_drift_allowed = 2 部署监控节点
# ceph-deploymon create-initial
found configuration file at: /root/.cephdeploy.conf
Invoked (1.5.36): /usr/bin/ceph-deploy mon create-initial
ceph-deploy options:
username : None
verbose : False
overwrite_conf : False
subcommand : create-initial
quiet : False
cd_conf :
cluster : ceph
func :
ceph_conf : None
default_release : False
keyrings : None
Deploying mon, cluster ceph hosts mon-1 osd-1 osd-2
......省略
Storing ceph.client.admin.keyring
Storing ceph.bootstrap-mds.keyring
keyring 'ceph.mon.keyring' already exists
Storing ceph.bootstrap-osd.keyring
Storing ceph.bootstrap-rgw.keyring
Destroy temp directory /tmp/tmpVLiIFr 目录内容如下
# ll
total 76
-rw------- 1 root root 113 Dec6 22:49 ceph.bootstrap-mds.keyring
-rw------- 1 root root 113 Dec6 22:49 ceph.bootstrap-osd.keyring
-rw------- 1 root root 113 Dec6 22:49 ceph.bootstrap-rgw.keyring
-rw------- 1 root root 129 Dec6 22:49 ceph.client.admin.keyring
-rw-r--r-- 1 root root 300 Dec6 22:47 ceph.conf
-rw-r--r-- 1 root root 50531 Dec6 22:49 ceph-deploy-ceph.log
-rw------- 1 root root 73 Dec6 22:46 ceph.mon.keyring 查看集群状态
# ceph -s
cluster 1b27aaf2-8b29-49b1-b50e-7ccb1f72d1fa
health HEALTH_ERR
no osds
monmap e1: 1 mons at {mon-1=192.168.50.123:6789/0}
election epoch 3, quorum 0 mon-1
osdmap e1: 0 osds: 0 up, 0 in
flags sortbitwise
pgmap v2: 64 pgs, 1 pools, 0 bytes data, 0 objects
0 kB used, 0 kB / 0 kB avail
64 creating 开始部署OSD：
# ceph-deploy--overwrite-conf osd preparemon-1:/dev/sdbmon-1:/dev/sdcmon-1:/dev/sdd osd-1:/dev/sdbosd-1:/dev/sdcosd-1:/dev/sddosd-2:/dev/sdbosd-2:/dev/sdcosd-2:/dev/sdd--zap-disk
这里如果部署osd-2有问题，就删除程序和目录，重新分区以后成功创建osd-2
ceph-deploy--overwrite-conf osd prepareosd-2:/dev/sdbosd-2:/dev/sdcosd-2:/dev/sdd# lsblk
NAME MAJ:MIN RMSIZE RO TYPE MOUNTPOINT
fd0 2:0 1 4K0 disk
sda 8:0 0 20G0 disk
├─sda1 8:1 0500M0 part /boot
└─sda2 8:2 0 19.5G0 part
├─centos-root 253:0 0 18.5G0 lvm/
└─centos-swap 253:1 0 1G0 lvm
sdb 8:16 0 2T0 disk
├─sdb1 8:17 0 2T0 part /var/lib/ceph/osd/ceph-0
└─sdb2 8:18 0 5G0 part
sdc 8:32 0 2T0 disk
├─sdc1 8:33 0 2T0 part /var/lib/ceph/osd/ceph-1
└─sdc2 8:34 0 5G0 part
sdd 8:48 0 2T0 disk
├─sdd1 8:49 0 2T0 part /var/lib/ceph/osd/ceph-2
└─sdd2 8:50 0 5G0 part
sr0 11:0 1603M0 rom
# lsblk
NAME MAJ:MIN RMSIZE RO TYPE MOUNTPOINT
fd0 2:0 1 4K0 disk
sda 8:0 0 20G0 disk
├─sda1 8:1 0500M0 part /boot
└─sda2 8:2 0 19.5G0 part
├─centos-root 253:0 0 18.5G0 lvm/
└─centos-swap 253:1 0 1G0 lvm
sdb 8:16 0 2T0 disk
├─sdb1 8:17 0 2T0 part /var/lib/ceph/osd/ceph-3
└─sdb2 8:18 0 5G0 part
sdc 8:32 0 2T0 disk
├─sdc1 8:33 0 2T0 part /var/lib/ceph/osd/ceph-4
└─sdc2 8:34 0 5G0 part
sdd 8:48 0 2T0 disk
├─sdd1 8:49 0 2T0 part /var/lib/ceph/osd/ceph-5
└─sdd2 8:50 0 5G0 part
sr0 11:0 1603M0 rom
# lsblk
NAME MAJ:MIN RMSIZE RO TYPE MOUNTPOINT
fd0 2:0 1 4K0 disk
sda 8:0 0 20G0 disk
├─sda1 8:1 0500M0 part /boot
└─sda2 8:2 0 19.5G0 part
├─centos-root 253:0 0 18.5G0 lvm/
└─centos-swap 253:1 0 1G0 lvm
sdb 8:16 0 2T0 disk
├─sdb1 8:17 0 2T0 part /var/lib/ceph/osd/ceph-6
└─sdb2 8:18 0 5G0 part
sdc 8:32 0 2T0 disk
├─sdc1 8:33 0 2T0 part /var/lib/ceph/osd/ceph-7
└─sdc2 8:34 0 5G0 part
sdd 8:48 0 2T0 disk
├─sdd1 8:49 0 2T0 part /var/lib/ceph/osd/ceph-8
└─sdd2 8:50 0 5G0 part
sr0 11:0 1603M0 rom
这里有个WARN，去掉这个WARN只需要增加rbd池的PG就好
# ceph osd pool set rbd pg_num 128
set pool 0 pg_num to 128
# ceph osd pool set rbd pgp_num 128
set pool 0 pgp_num to 128
# ceph -s
cluster 0865fe85-1655-4208-bed6-274cae945746
health HEALTH_OK
monmap e3: 2 mons at {mon-1=192.168.50.123:6789/0,osd-1=192.168.50.124:6789/0}
election epoch 30, quorum 0,1 mon-1,osd-1
osdmap e58: 9 osds: 9 up, 9 in
flags sortbitwise
pgmap v161: 128 pgs, 1 pools, 0 bytes data, 0 objects
310 MB used, 18377 GB / 18378 GB avail
128 active+clean 给各个节点推送config配置
请不要直接修改某个节点的/etc/ceph/ceph.conf文件的，而是要去部署节点(此处为ceph-1:/root/cluster/ceph.conf)目录下修改。因为节点到几十个的时候，不可能一个个去修改的，采用推送的方式快捷安全！
修改完毕后，执行如下指令，将conf文件推送至各个节点：
# ceph-deploy --overwrite-conf config push mon-1 osd-1 osd-2
found configuration file at: /root/.cephdeploy.conf
Invoked (1.5.36): /usr/bin/ceph-deploy admin mon-1 osd-1 osd-2
ceph-deploy options:
username : None
verbose : False
overwrite_conf : False
quiet : False
cd_conf :
cluster : ceph
client : ['mon-1', 'osd-1', 'osd-2']
func :
ceph_conf : None
default_release : False
Pushing admin keys and conf to mon-1
connected to host: mon-1
detect platform information from remote host
detect machine type
write cluster configuration to /etc/ceph/{cluster}.conf
# ceph-deploy --overwrite-conf config push ceph-1 ceph-2 ceph-3
Pushing admin keys and conf to osd-1
connected to host: osd-1
detect platform information from remote host
detect machine type
write cluster configuration to /etc/ceph/{cluster}.conf
Pushing admin keys and conf to osd-2
connected to host: osd-2
detect platform information from remote host
detect machine type
write cluster configuration to /etc/ceph/{cluster}.conf
此时，需要重启各个节点的monitor服务
mon和osd的启动方式
#mon-1为各个monitor所在节点的主机名。
systemctl start ceph-mon@mon-1.service
systemctl restart ceph-mon@mon-1.service
systemctl stop ceph-mon@mon-1.service
#0为该节点的OSD的id，可以通过`ceph osd tree`查看
systemctl start/stop/restart ceph-osd@0.service
遇到的报错
1. Monitor clock skew detected
http://www.iyunv.com/thread-157755-1-1.html
1. Monitor clock skew detected
# ceph -s
cluster f25ad2c5-fd2a-4fcc-a522-344eb498fee5
health HEALTH_ERR
clock skew detected on mon.osd-2
64 pgs are stuck inactive for more than 300 seconds
64 pgs stuck inactive
no osds
Monitor clock skew detected 添加配置参数：
vim /etc/ceph/ceph.conf
mon clock drift allowed = 2
mon clock drift warn backoff = 30
同步配置文件
ceph-deploy --overwrite-conf admin osd-1 osd-2 重启mon服务
systemctl restart ceph-mon@osd-2.service 问题总结：
本问题主要是mon节点服务器，时间偏差比较大导致，本次遇到问题为测试环境，通过修改ceph对时间偏差阀值，规避的告警信息，线上业务环境，注意排查服务器时间同步问题。
参考网址：
http://xuxiaopang.com/2016/10/09/ceph-quick-install-el7-jewel/
http://bbs.ceph.org.cn/question/205
http://www.docoreos.com/?p=86
http://www.voidcn.com/blog/xiangpingli/article/p-5979182.html
页:
[1]