如何使用ceph當作nova backend storage

在ceph node底下

# 建立ceph pool 給nova使用,pool名稱為vms $ sudo ceph osd pool create vms 128

建立一個user為client.nova並對pool vms相關權限設定,產生一個keyring file

$ ceph auth get-or-create client.nova mon 'allow r' osd 'allow class-read object_prefix rbd_children, allow rwx pool=vms, allow rx pool=images' -o /etc/ceph/ceph.client.nova.keyring # 複製ceph.conf與nova.keyring到openstack controller上 $ scp /etc/ceph/ceph.conf root@openstack node ip:/etc/ceph $ scp /etc/ceph/ceph.client.nova.keyring root@openstack node ip:/etc/ceph # 複製client.nova的key複製到controller /tmp/client.cinder.key上 $ ceph auth get-key client.nova | ssh openstack node ip tee /tmp/client.nova.key

添加一个secret key到libvirt

$ uuidgen 7ad56fa3-23e2-4d86-ae9e-253326a704f5 $ cat > secret.xml <<EOF <secret ephemeral='no' private='no'> <uuid>7ad56fa3-23e2-4d86-ae9e-253326a704f5</uuid> <usage type='ceph'> <name>client.nova secret</name> </usage> </secret> EOF $ virsh secret-define --file secret.xml Secret 7ad56fa3-23e2-4d86-ae9e-253326a704f5 created $ virsh secret-set-value --secret 7ad56fa3-23e2-4d86-ae9e-253326a704f5 $(cat /tmp/client.nova.key) # 查詢密鑰 $ virsh secret-list  UUID                                  Usage ------------------------------------------------------------- d65bc749-e160-44d3-a903-9c50291b4293  ceph client.cinder secret 7ad56fa3-23e2-4d86-ae9e-253326a704f5 ceph client.nova secret

我們這邊的nova-compute服務是跑在openstack controller上

# 在opentack controller上的ceph.conf新增cinder的setion $ vim /etc/ceph/ceph.conf [client.nova] keyring = /etc/ceph/ceph.client.nova.keyring # 修改權限 $ chmod 0640 /etc/ceph/ceph.client.nova.keyring $ chown nova:nova /etc/ceph/ceph.client.nova.keyring # 修改/etc/nova/nova-compute.conf $ sudo cp /etc/nova/nova-compute.conf /etc/nova/nova-compute.conf.bak $ sudo vim /etc/nova/nova-compute.conf # 修改或是新增以下參數 [libvirt] virt_type=qemu images_rbd_ceph_conf = /etc/ceph/ceph.conf images_rbd_pool = vms images_type = rbd rbd_user = nova rbd_secret_uuid = 7ad56fa3-23e2-4d86-ae9e-253326a704f5 disk_cachemodes="network=writeback" # 重啟 nova-compute服務 $ restart nova-compute

驗證是否連接ceph成功

$ nova boot --flavor m1.tiny --image image-name instance1

參考來源:

Ceph与OpenStack的Nova相结合