视频： http://b23.tv/lGqU9OJ
| 主机名 | IP | 内存 | CPU |
|---|---|---|---|
| hci-cloud01 | 10.10.10.60，20.20.20.60，30.30.30.60 | 32G | 12核 |
| hci-cloud02 | 10.10.10.61，20.20.20.61，30.30.30.61 | 32G | 12核 |
| hci-cloud03 | 10.10.10.62，20.20.20.62，30.30.30.62 | 32G | 12核 |
每台机器磁盘最少需要3块裸盘做ceph
链接：https://pan.baidu.com/s/1IDy5WLq1JHhVoMXPU55jjg?pwd=6666
提取码：6666
sed -i "s/SELINUX=enforcing/SELINUX=edisabled/g" /etc/selinux/config; systemctl stop firewalld; systemctl disable firewalld; setenforce 0
ops-director-kolla-new.tar.gz
tar-tool.rpm.xz
xz -d tar-tool.rpm.xz
rpm -ivh tar-tool.rpm
tar xvf ops-director-kolla-new.tar.gz
ssh-keygen
一直回车即可
for i in 0 1 2;do ssh-copy-id -i /root/.ssh/id_rsa.pub [email protected]$i;done
yes
主机密码
cd ops-director-kolla
vi director-inventory.ini
1 [all:vars]
3 ansible_ssh_pass=000000
[deployment]
6 10.10.10.60
[control]
9 10.10.10.60 hostname=hci-cloud01
10 10.10.10.61 hostname=hci-cloud02
11 10.10.10.62 hostname=hci-cloud03
[compute]
14 10.10.10.60
15 10.10.10.61
16 10.10.10.62
vi start.sh
# 管理实例网卡
7 Openstack_Tenant_NIC="eth2"
# 执行脚本
source start.sh
exit
再连接即可(ssh [email protected]/1/2)
cd ops-director-kolla/ceph-director/
vim inventory/ceph-cluster.ini
2 [all:vars]
4 ansible_ssh_pass=000000
7 [mons]
8 10.10.10.60
9 10.10.10.61
10 10.10.10.62
11
12 [osds]
13 10.10.10.60
14 10.10.10.61
15 10.10.10.62
PS: 如果存在不同的磁盘名称情况下，指定磁盘操作如下，如果都一样无需在这操作
12 [osds]
13 10.10.10.60 devices="['/dev/sdb','/dev/sdc','/dev/sdd']"
14 10.10.10.61 devices="['/dev/sde','/dev/sdf','/dev/sdg']"
15 10.10.10.62 devices="['/dev/sdj','/dev/sdk','/dev/sdl']"
cd ops-director-kolla/ceph-director/
vim group_vars/all.yml
158 yum_repo_ip: 10.10.10.60
329 monitor_interface: eth0
370 public_network: "10.10.10.0/24"
371 cluster_network: "20.20.20.0/24"
408 radosgw_interface: eth0
577 ceph_docker_registry: 10.10.10.60:8086
848 insecure_registries: "10.10.10.60:8086"
vim group_vars/osds.yml
36 devices:
37 - /dev/sdb
38 - /dev/sdc
39 - /dev/sdd
40 - /dev/sde
PS: 注意根据自身环境写入，必须是裸盘，而且每台集群磁盘名称一致（磁盘查看命令: lsblk）
mds配置
```shell vim ceph-director/group_vars/mdss.yml
根据实际服务器配置内存和CPU,可使用默认配置不修改
ceph_mds_docker_memory_limit: "4096m" ceph_mds_docker_cpu_limit: 4 ```
mgr配置
```shell vim ceph-director/group_vars/mgrs.yml
根据实际服务器配置内存和CPU,可使用默认配置不修改
ceph_mgr_docker_memory_limit: "4096m" ceph_mgr_docker_cpu_limit: 1 ```
mon配置
```shell vim ceph-director/group_vars/mons.yml
根据实际服务器配置内存和CPU,可使用默认配置不修改
ceph_mon_docker_memory_limit: "4096m" ceph_mon_docker_cpu_limit: 1 ```
rgws配置
```shell vim ceph-director/group_vars/rgws.yml
根据实际服务器配置内存和CPU,可使用默认配置不修改
ceph_rgw_docker_memory_limit: "4096m" ceph_rgw_docker_cpu_limit: 8 ```
cd /root/ops-director-kolla/ceph-director
source ansible-playbook-command-docker.sh
# 查看容器ID
docker ps | grep ceph
docker exec -it ceph任意容器id bash
#关闭mon监测
ceph config set mon auth_allow_insecure_global_id_reclaim false
查看ceph监控服务
ceph mgr services
访问：
https://10.10.10.60:8443/
用户名：admin
密码：p@ssw0rd
# 查看
cat /root/ops-director-kolla/ceph-director/group_vars/all.yml | grep dashboard_admin_password
http://10.10.10.60:9283/
cd /root/ops-director-kolla
vim inventory/multinode
5 ansible_ssh_pass=000000
vim /etc/kolla/globals.yml
5 kolla_internal_vip_address: "10.10.10.248"
6 docker_registry: "10.10.10.60:8086"
8 network_interface: "eth0"
9 neutron_external_interface: "eth2"
# 如果是服务器那么就使用kvm
35 nova_compute_virt_type: "qemu"
source ops-install.sh
http://10.10.10.248/
用户名：root
密码：cat /etc/kolla/passwords.yml | grep keystone_admin_password
source /etc/kolla/admin-openrc.sh
openstack image create cirros04 --disk-format qcow2 --file cirros-0.4.0-x86_64-disk.img
docker exec -it ceph-mon-hci-cloud01 bash
# 得到一个id，可以查看镜像id是否能匹配上
rbd ls images
openstack volume type create Ceph --public --description "Cephcinder"
openstack volume create cinder01 --size 50
docker exec -it ceph-mon-hci-cloud01 bash
# 得到一个id，可以查看镜像id是否能匹配上
rbd ls volumes
source /etc/kolla/admin-openrc.sh
openstack service create --name=swift --description="Swift Service" object-store
openstack user create --domain default --password swift swift
openstack role add --project service --user swift admin
openstack endpoint create swift internal "http://10.10.10.60:8080/swift/v1" --region RegionOne
openstack endpoint create swift public "http://10.10.10.60:8080/swift/v1" --region RegionOne
openstack endpoint create swift admin "http://10.10.10.60:8080/swift/v1" --region RegionOne
vim /etc/ceph/ceph.conf
26 [client]
27 rbd_default_features = 1
28 rgw_dynamic_resharding = False
29 rgw_enable_ops_log = False
30 rgw_enable_usage_log = True
31 rgw_keystone_accepted_admin_roles = admin
32 rgw_keystone_accepted_roles = admin,_member_,member
33 rgw_keystone_admin_domain = Default
# 密码在/etc/kolla/passwords.yml
34 rgw_keystone_admin_password = Kn2e75mEKQDMChTfPnOkFKBtU9bkIu3YzMkeZTPb
35 rgw_keystone_admin_project = admin
36 rgw_keystone_admin_user = admin
37 rgw_keystone_api_version = 3
38 rgw_keystone_token_cache_size = 500
39 rgw_keystone_revocation_interval = 0
40 # openstack_vip
41 rgw_keystone_url = http://10.10.10.248:5000
42 rgw_s3_auth_use_keystone = True
43 rgw_user_quota_bucket_sync_interval = 60
44 rgw_user_quota_sync_interval = 3600
docker restart <rgw_container>
openstack container create ceph_swfit --public
docker exec -it ceph-mon-hci-cloud01 bash
radosgw-admin bucket list
# 查看网络名称
cat /etc/kolla/neutron-server/ml2_conf.ini
查看实例后端
docker exec -it ceph-mon-hci-cloud01 bash
rbd ls vms