一、基本环境介绍：
本文采用ceph-deploy工具进行ceph安装，ceph-deploy可以单独作为一个admin节点，也可以装在任意node节点上。
系统环境如下：
1、 系统采用redhat-6.5x86_64 basic_server安装，总共3个节点,ntp时间同步
2、 关闭selinux，用epel及ceph官方源，版本为0.86
3、 3个节点已做互信，并配置host，每个节点上有3个磁盘做osd
4、 升级内核为3.18，并从新内核启动。
二、安装步骤：
1、 每个node上设置iptables规则或关闭iptables(eth0指ceph网络所在的网卡名称)
|
1
2
3
|
iptables -A INPUT -i eth0 -p tcp -s0.0.0.0
/0
--dport 6789 -j ACCEPT
iptables -A INPUT -i eth0 -m multiport -ptcp -s 0.0.0.0
/0
--dports 6800:7300 -j ACCEPT
service iptables save
|
2、 格式化、挂载osd
|
1
2
3
4
5
|
yum -y
install
xfsprogs
mkfs.xfs
/dev/sdb
mkdir
/osd
{0..2}
#blkid 查看sdb的UUID
echo
'UUID=89048e27-ff01-4365-a103-22e95fb2cc93/osd0 xfs noatime,nobarrier,nodiratime 00'
>>
/etc/fstab
|
一个磁盘对应一个osd，每个节点都创建osd0、osd1、osd2目录，对应的磁盘挂载相应目录即可。
3、安装ceph部署工具
|
1
2
3
|
#mkdir ceph #最好创建一个目录，因为安装ceph过程中，会在安装目录生成一些文件
#cd ceph
#yum -y install ceph-deploy
|
4、创建mon
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
|
ceph-deploy new node1 nod2 node3
#这个命令其实就是仅仅生成了ceph.conf和ceph.mon.keyring 文件
vim ceph.conf 追加以下内容(根据需求更改)
debug_ms = 0
mon_clock_drift_allowed = 1
osd_pool_default_size = 2
#副本数量
osd_pool_default_min_size = 1
osd_pool_default_pg_num = 128
#pg数量
osd_pool_default_pgp_num = 128
osd_crush_chooseleaf_type = 0
debug_auth = 0
/0
debug_optracker = 0
/0
debug_monc = 0
/0
debug_crush = 0
/0
debug_buffer = 0
/0
debug_tp = 0
/0
debug_journaler = 0
/0
debug_journal = 0
/0
debug_lockdep = 0
/0
debug_objclass = 0
/0
debug_perfcounter = 0
/0
debug_timer = 0
/0
debug_filestore = 0
/0
debug_context = 0
/0
debug_finisher = 0
/0
debug_heartbeatmap = 0
/0
debug_asok = 0
/0
debug_throttle = 0
/0
debug_osd = 0
/0
debug_rgw = 0
/0
debug_mon = 0
/0
osd_max_backfills = 4
filestore_split_multiple = 8
filestore_fd_cache_size = 1024
filestore_queue_committing_max_bytes =1048576000
filestore_queue_max_ops = 500000
filestore_queue_max_bytes = 1048576000
filestore_queue_committing_max_ops = 500000
osd_max_pg_log_entries = 100000
osd_mon_heartbeat_interval = 30
#Performance tuning filestore
osd_mount_options_xfs =rw,noatime,logbsize=256k,delaylog
#osd_journal_size = 20480 日志大小，不指定，默认是5G
osd_op_log_threshold = 50
osd_min_pg_log_entries = 30000
osd_recovery_op_priority = 1
osd_mkfs_options_xfs = -f -i size=2048
osd_mkfs_type = xfs
osd_journal =
/var/lib/ceph/osd/
$cluster-$
id
/journal
journal_queue_max_ops = 500000
journal_max_write_bytes = 1048576000
journal_max_write_entries = 100000
journal_queue_max_bytes = 1048576000
objecter_infilght_op_bytes = 1048576000
objecter_inflight_ops = 819200
ms_dispatch_throttle_bytes = 1048576000
sd_data =
/var/lib/ceph/osd/
$cluster-$
id
merge_threshold = 40
backfills = 1
mon_osd_min_down_reporters = 13
mon_osd_down_out_interval = 600
rbd_cache_max_dirty_object = 0
rbd_cache_target_dirty = 235544320
rbd_cache_writethrough_until_flush =
false
rbd_cache_size = 335544320
rbd_cache_max_dirty = 335544320
rbd_cache_max_dirty_age = 60
rbd_cache =
false
|
5、安装ceph
#On All nodes to Install
|
1
|
yum -y
install
ceph
|
admin节点上执行：
|
1
2
|
ceph-deploy mon create node1 node2 node3
ceph-deploy gatherkeys node1
#从moniter节点获得keys，用来管理节点
|
6、创建激活osd
|
1
2
3
4
5
|
ceph-deploy osd prepare node1:
/osd0node1
:
/osd1
node1:
/osd2
node2:
/osd0
node2:
/osd1
node2:
/osd2
node3:
/osd0node3
:
/osd1
node3:
/osd2
ceph-deploy osd activate node1:
/osd0node1
:
/osd1
node1:
/osd2
node2:
/osd0
node2:
/osd1
node2:
/osd2
node3:
/osd0node3
:
/osd1
node3:
/osd2
ceph-deploy admin node1 node2 node3
#从admin节点复制配置文件及key到node
chmod
+r
/etc/ceph/ceph
.client.admin.keyring (所有节点为该文件添加读权限)
|
也可以创建相关的pool
|
1
2
3
4
|
ceph osd pool create volumes 128
ceph osd pool create images 128
ceph osd pool create vms 128
ceph osd pool create backups 128
|
如果ceph安装失败，要清除ceph环境，执行以下命令
|
1
2
3
|
ceph-deploy purge node1 node2 node3 (所有节点)
ceph-deploy purgedata node1 node2 node3 (所有节点)
ceph-deploy forgetkeys
|
本文转自Jacken_yang 51CTO博客，原文链接：http://blog.51cto.com/linuxnote/1788333，如需转载请自行联系原作者
