一、实施步骤：
（1）客户端也要有cent用户：
useradd cent && echo "123" | passwd --stdin cent
echo -e 'Defaults:cent !requiretty\ncent ALL = (root) NOPASSWD:ALL' | tee /etc/sudoers.d/ceph
chmod 440 /etc/sudoers.d/ceph
（2）openstack要用ceph的节点（比如compute-node和storage-node）安装下载的软件包：
yum localinstall ./* -y
或则：每个节点安装 clients（要访问ceph集群的节点）：
yum install python-rbdyum
install ceph-common
如果先采用上面的方式安装客户端，其实这两个包在rpm包中早已经安装过了
（3）部署节点上执行，为openstack节点安装ceph：
ceph-deploy install controller
ceph-deploy admin controller
（4）客户端执行
sudo chmod 644 /etc/ceph/ceph.client.admin.keyring| 1 |
（5）create pools，只需在一个ceph节点上操作即可：
ceph osd pool create images 1024
ceph osd pool create vms 1024
ceph osd pool create volumes 1024
显示pool的状态
ceph osd lspools
（6）在ceph集群中，创建glance和cinder用户， 只需在一个ceph节点上操作即可：
ceph auth get-or-create client.glance mon 'allow r' osd 'allow class-read object_prefix rbd_children, allow rwx pool=images'
ceph auth get-or-create client.cinder mon 'allow r' osd 'allow class-read object_prefix rbd_children, allow rwx pool=volumes, allow rwx pool=vms, allow rx pool=images'
nova使用cinder用户，就不单独创建了
（7）拷贝ceph-ring， 只需在一个ceph节点上操作即可：
| 1 2 | ceph auth get-or-create client.glance > /etc/ceph/ceph.client.glance.keyring ceph auth get-or-create client.cinder > /etc/ceph/ceph.client.cinder.keyring |
使用scp拷贝到其他节点（ceph集群节点和openstack的要用ceph的节点比如compute-node和storage-node，本次对接的是一个all-in-one的环境，所以copy到controller节点即可 ）
| 1 2 3 4 | [root@yunwei ceph]# ls ceph.client.admin.keyring ceph.client.cinder.keyring ceph.client.glance.keyring ceph.conf rbdmap tmpR3uL7W [root@yunwei ceph]# [root@yunwei ceph]# scp ceph.client.glance.keyring ceph.client.cinder.keyring controller:/etc/ceph/ |
（8）更改文件的权限(所有客户端节点均执行)
| 1 2 | chown glance:glance /etc/ceph/ceph.client.glance.keyring chown cinder:cinder /etc/ceph/ceph.client.cinder.keyring |
（9）更改libvirt权限（只需在nova-compute节点上操作即可，每个计算节点都做）
uuidgen
940f0485-e206-4b49-b878-dcd0cb9c70a4
在/etc/ceph/目录下（在什么目录没有影响，放到/etc/ceph目录方便管理）：
cat > secret.xml <<EOF
<secret ephemeral='no' private='no'>
<uuid>940f0485-e206-4b49-b878-dcd0cb9c70a4</uuid>
<usage type='ceph'>
<name>client.cinder secret</name>
</usage>
</secret>
EOF
将 secret.xml 拷贝到所有compute节点，并执行：：
virsh secret-define --file secret.xml
ceph auth get-key client.cinder > ./client.cinder.key
virsh secret-set-value --secret 940f0485-e206-4b49-b878-dcd0cb9c70a4 --base64 $(cat ./client.cinder.key)配置Glance， 在所有的controller节点上做如下更改：
vim /etc/glance/glance-api.conf
[DEFAULT]
default_store = rbd
[cors]
[cors.subdomain]
[database]
connection = mysql+pymysql://glance:GLANCE_DBPASS@controller/glance
[glance_store]
stores = rbd
default_store = rbd
rbd_store_pool = images
rbd_store_user = glance
rbd_store_ceph_conf = /etc/ceph/ceph.conf
rbd_store_chunk_size = 8
[image_format]
[keystone_authtoken]
auth_uri = http://controller:5000
auth_url = http://controller:35357
memcached_servers = controller:11211
auth_type = password
project_domain_name = default
user_domain_name = default
project_name = service
username = glance
password = glance
[matchmaker_redis]
[oslo_concurrency]
[oslo_messaging_amqp]
[oslo_messaging_kafka]
[oslo_messaging_notifications]
[oslo_messaging_rabbit]
[oslo_messaging_zmq]
[oslo_middleware]
[oslo_policy]
[paste_deploy]
flavor = keystone
[profiler]
[store_type_location_strategy]
[task]
[taskflow_executor]
在所有的controller节点上做如下更改
| 1 2 | systemctl restart openstack-glance-api.service systemctl status openstack-glance-api.service |
创建image验证：
| 1 2 3 4 5 | [root@controller ~]# openstack image create "cirros" --file cirros-0.3.3-x86_64-disk.img.img --disk-format qcow2 --container-format bare --public
[root@controller ~]# rbd ls images 9ce5055e-4217-44b4-a237-e7b577a20dac **********有输出镜像说明成功了 |
（8）配置 Cinder：
vim /etc/cinder/cinder.conf
[DEFAULT]
my_ip = 172.16.254.63
glance_api_servers = http://controller:9292
auth_strategy = keystone
enabled_backends = ceph
state_path = /var/lib/cinder
transport_url = rabbit://openstack:admin@controller
[backend]
[barbican]
[brcd_fabric_example]
[cisco_fabric_example]
[coordination]
[cors]
[cors.subdomain]
[database]
connection = mysql+pymysql://cinder:CINDER_DBPASS@controller/cinder
[fc-zone-manager]
[healthcheck]
[key_manager]
[keystone_authtoken]
auth_uri = http://controller:5000
auth_url = http://controller:35357
memcached_servers = controller:11211
auth_type = password
project_domain_name = default
user_domain_name = default
project_name = service
username = cinder
password = cinder
[matchmaker_redis]
[oslo_concurrency]
lock_path = /var/lib/cinder/tmp
[oslo_messaging_amqp]
[oslo_messaging_kafka]
[oslo_messaging_notifications]
[oslo_messaging_rabbit]
[oslo_messaging_zmq]
[oslo_middleware]
[oslo_policy]
[oslo_reports]
[oslo_versionedobjects]
[profiler]
[ssl]
[ceph]
volume_driver = cinder.volume.drivers.rbd.RBDDriver
rbd_pool = volumes
rbd_ceph_conf = /etc/ceph/ceph.conf
rbd_flatten_volume_from_snapshot = false
rbd_max_clone_depth = 5
rbd_store_chunk_size = 4
rados_connect_timeout = -1
glance_api_version = 2
rbd_user = cinder
rbd_secret_uuid = 940f0485-e206-4b49-b878-dcd0cb9c70a4
volume_backend_name=ceph
重启cinder服务：
| 1 2 | systemctl restart openstack-cinder-api.service openstack-cinder-scheduler.service openstack-cinder-volume.service systemctl status openstack-cinder-api.service openstack-cinder-scheduler.service openstack-cinder-volume.service |
创建volume验证：
| 1 2 | [root@controller gfs]# rbd ls volumes volume-43b7c31d-a773-4604-8e4a-9ed78ec18996 |
（9）配置Nova：
vim /etc/nova/nova.conf
[DEFAULT]
my_ip=172.16.254.63
use_neutron = True
firewall_driver = nova.virt.firewall.NoopFirewallDriver
enabled_apis=osapi_compute,metadata
transport_url = rabbit://openstack:admin@controller
[api]
auth_strategy = keystone
[api_database]
connection = mysql+pymysql://nova:NOVA_DBPASS@controller/nova_api
[barbican]
[cache]
[cells]
[cinder]
os_region_name = RegionOne
[cloudpipe]
[conductor]
[console]
[consoleauth]
[cors]
[cors.subdomain]
[crypto]
[database]
connection = mysql+pymysql://nova:NOVA_DBPASS@controller/nova
[ephemeral_storage_encryption]
[filter_scheduler]
[glance]
api_servers = http://controller:9292
[guestfs]
[healthcheck]
[hyperv]
[image_file_url]
[ironic]
[key_manager]
[keystone_authtoken]
auth_uri = http://controller:5000
auth_url = http://controller:35357
memcached_servers = controller:11211
auth_type = password
project_domain_name = default
user_domain_name = default
project_name = service
username = nova
password = nova
[libvirt]
virt_type=qemu
images_type = rbd
images_rbd_pool = vms
images_rbd_ceph_conf = /etc/ceph/ceph.conf
rbd_user = cinder
rbd_secret_uuid = 940f0485-e206-4b49-b878-dcd0cb9c70a4
[matchmaker_redis]
[metrics]
[mks]
[neutron]
url = http://controller:9696
auth_url = http://controller:35357
auth_type = password
project_domain_name = default
user_domain_name = default
region_name = RegionOne
project_name = service
username = neutron
password = neutron
service_metadata_proxy = true
metadata_proxy_shared_secret = METADATA_SECRET
[notifications]
[osapi_v21]
[oslo_concurrency]
lock_path=/var/lib/nova/tmp
[oslo_messaging_amqp]
[oslo_messaging_kafka]
[oslo_messaging_notifications]
[oslo_messaging_rabbit]
[oslo_messaging_zmq]
[oslo_middleware]
[oslo_policy]
[pci]
[placement]
os_region_name = RegionOne
auth_type = password
auth_url = http://controller:35357/v3
project_name = service
project_domain_name = Default
username = placement
password = placement
user_domain_name = Default
[quota]
[rdp]
[remote_debug]
[scheduler]
[serial_console]
[service_user]
[spice]
[ssl]
[trusted_computing]
[upgrade_levels]
[vendordata_dynamic_auth]
[vmware]
[vnc]
enabled=true
vncserver_listen=$my_ip
vncserver_proxyclient_address=$my_ip
novncproxy_base_url = http://172.16.254.63:6080/vnc_auto.html
[workarounds]
[wsgi]
[xenserver]
[xvp]
重启nova服务：
| 1 2 | systemctl restart openstack-nova-api.service openstack-nova-consoleauth.service openstack-nova-scheduler.service openstack-nova-compute.service openstack-nova-cert.service systemctl status openstack-nova-api.service openstack-nova-consoleauth.service openstack-nova-scheduler.service openstack-no |
ceph常用命令
1、查看ceph集群配置信息
| 1 | ceph daemon /var/run/ceph/ceph-mon.$(hostname -s).asok config show |
2、在部署节点修改了ceph.conf文件，将新配置推送至全部的ceph节点
| 1 | ceph-deploy --overwrite-conf config push dlp node1 node2 node3 |
3、检查仲裁状态，查看mon添加是否成功
| 1 | ceph quorum_status --format json-pretty |
4、列式pool列表
| 1 | ceph osd lspools |
5、列示pool详细信息
| 1 | ceph osd dump |grep pool |
6、检查pool的副本数
| 1 | ceph osd dump|grep -i size |
7、创建pool
| 1 | ceph osd pool create pooltest 128 |
8、删除pool
| 1 2 | ceph osd pool delete data ceph osd pool delete data data --yes-i-really-really-mean-it |
9、设置pool副本数
| 1 2 | ceph osd pool get data size ceph osd pool set data size 3 |
10、设置pool配额
| 1 2 | ceph osd pool set-quota data max_objects 100 #最大100个对象 ceph osd pool set-quota data max_bytes $((10 * 1024 * 1024 * 1024)) #容量大小最大为10G |
11、重命名pool
| 1 | ceph osd pool rename data date |
12、PG, Placement Groups。CRUSH先将数据分解成一组对象，然后根据对象名称、复制级别和系统中的PG数等信息执行散列操作，再将结果生成PG ID。可以将PG看做一个逻辑容器，这个容器包含多个对 象，同时这个逻辑对象映射之多个OSD上。如果没有PG，在成千上万个OSD上管理和跟踪数百万计的对象的复制和传播是相当困难的。没有PG这一层，管理海量的对象所消耗的计算资源也是不可想象的。建议每个OSD上配置50~100个PG。
PGP是为了实现定位而设置的PG，它的值应该和PG的总数(即pg_num)保持一致。对于Ceph的一个pool而言，如果增加pg_num，还应该调整pgp_num为同样的值，这样集群才可以开始再平衡。
参数pg_num定义了PG的数量，PG映射至OSD。当任意pool的PG数增加时，PG依然保持和源OSD的映射。直至目前，Ceph还未开始再平衡。此时，增加pgp_num的值，PG才开始从源OSD迁移至其他的OSD，正式开始再平衡。PGP，Placement Groups of Placement。
计算PG数：
ceph集群中的PG总数
| 1 | PG总数 = (OSD总数 * 100) / 最大副本数 ** 结果必须舍入到最接近的2的N次方幂的值 |
ceph集群中每个pool中的PG总数
| 1 | 存储池PG总数 = (OSD总数 * 100 / 最大副本数) / 池数 |
获取现有的PG数和PGP数值
| 1 2 | ceph osd pool get data pg_num ceph osd pool get data pgp_num |
13、修改存储池的PG和PGP
| 1 2 | ceph osd pool set data pg_num = 1 ceph osd pool set data pgp_num = 1 |
