kolla 部署 ceph

来看看

　　一、环境准备
　　1、节点准备

server	组件	disk
deploy	安装kolla，用于部署	100G，osd
node0	ceph,osd	100G，osd
node1	ceph,osd	100G，osd
node2	ceph,osd	100G，osd

　　2、软件版本

kolla	newton
ceph	Jewel

　　3、配置主名
192.168.56.103 node0
192.168.56.104 node1
192.168.56.105 node2
　　4、准备image
　　参看： kolla newton 配置

　　5、在deploy上安装kolla
　　参看： kolla newton 配置
　　二、安装ceph
　　1、配置kolla ansible inventory

　　编辑/usr/share/kolla/ansible/inventory/multinode
...
[control]
node[0:2]
...
[storage]
node[0:2]
...　　注释掉，[network]、 [compute]、 [monitoring]下的所有主机
　　2、准备disk
ansible -i  /usr/share/kolla/ansible/inventory/multinode \
  all \
  -m shell  -a \
  'parted /dev/sdb -s -- mklabel gpt mkpart KOLLA_CEPH_OSD_BOOTSTRAP 1 -1'　　3、编辑/etc/kolla/globals.yml，注意替换node0_ip，docker registry的namespace，user，password

kolla_base_distro: "centos"
kolla_install_type: "source"
openstack_release: "3.0.0"

kolla_internal_vip_address: "node0_ip"
kolla_external_vip_address: "{{ kolla_internal_vip_address }}"
...
docker_registry: "registry_ip_address:registry_ip_port"
docker_namespace: "kolla"
docker_registry_username: "registry_username"
docker_registry_password: "registry_password"
...
enable_haproxy: "no"
enable_keystone: "no"
enable_glance: "no"
enable_neutron: "no"
enable_heat: "no"
enable_nova: "no"
enable_horizon: "no"
enable_central_logging: "no"
enable_ceph: "yes"
...
ceph_pool_type: "replicated"　　4、部署ceph
kolla-genpwd
kolla-ansible -i /usr/share/kolla/ansible/inventory/multinode prechecks
kolla-ansible -i /usr/share/kolla/ansible/inventory/multinode deploy 2>&1 |tee -a ceph.logs
kolla-ansible -i /usr/share/kolla/ansible/inventory/multinode post-deploy　　三、管理ceph
docker exec ceph_mon ceph -s