亲测100%可成功的openstack实例热迁移+冷迁移 openstack-Mitak（双节点）+centos7.2-1511
Description：最近实验好多次热迁移都是败了，网上的博客大多数都无法成功的进行热迁移，最后还是在老外的一片关于opensatck live migration 说明上找到的原因：[libvirtd] 没有配置 block_migration_flag= VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_NON_SHARED_INC
附上地址：http://remarks.peurat.net/rdo-live-migrationblock-migration/ （需要哦）
实验工具：
| openstack-Mitaka双节点（CentOS 7.2 1511版） | |
| host | IP |
| controller-node | 192.168.100.10 |
| compute-node1（控制节点上的计算服务） |
192.168.100.10 |
| compute-node2 | 192.168.100.20 |
一、首先看下热迁移，后期再写使用NFS,ceph等外部存储的实例热迁移；（本次热迁移不使用任何外部的存储，如NFS、ceph等。）：
1、保证两个节点的nova用户uid，gid相同，可使用id nova命令查看：
[root@compute ~]# id nova
uid=162(nova) gid=162(nova) groups=162(nova),99(nobody),107(qemu)
[root@controller ~]# id nova
uid=162(nova) gid=162(nova) groups=162(nova),99(nobody),107(qemu)
如果不相同着修改uid，gid：
# [root@compute ~]# usermod -u uid nova XXX
# [root@compute ~]# gropumod -g gid nova XXX2、修改libvirt配置文件（注意：两个节点都需要修改）：
[root@controller ~]# vi /etc/sysconfig/libvirtd
# Override the default config file
# NOTE: This setting is no longer honoured if using
# systemd. Set '--config /etc/libvirt/libvirtd.conf'
# in LIBVIRTD_ARGS instead.
LIBVIRTD_CONFIG=/etc/libvirt/libvirtd.conf
# Listen for TCP/IP connections
# NB. must setup TLS/SSL keys prior to using this
LIBVIRTD_ARGS="--listen"
[root@compute ~]# vi /etc/sysconfig/libvirtd
# Override the default config file
# NOTE: This setting is no longer honoured if using
# systemd. Set '--config /etc/libvirt/libvirtd.conf'
# in LIBVIRTD_ARGS instead.
LIBVIRTD_CONFIG=/etc/libvirt/libvirtd.conf
# Listen for TCP/IP connections
# NB. must setup TLS/SSL keys prior to using this
LIBVIRTD_ARGS="--listen"
[root@controller ~]# grep -vE '^#|^$' /etc/libvirt/libvirtd.conf
listen_tls = 0
listen_tcp = 1
auth_tcp = "none"
[root@compute ~]# grep -vE '^#|^$' /etc/libvirt/libvirtd.conf
listen_tls = 0
listen_tcp = 1
auth_tcp = "none"
两个节点都重启服务：systemctl restart libvirtd.service
两个节点都检查端口：ss -ntl | grep 16509
[root@compute ~]# ss -ntl | grep 16509
LISTEN 0 30 *:16509 *:*
[root@controller ~]# ss -ntl | grep 16509
LISTEN 0 30 *:16509 *:*
LISTEN 0 30 :::16509 :::*
再来测试下可否无密连接：
[root@controller ~]# virsh -c qemu+tcp://compute/system
Welcome to virsh, the virtualization interactive terminal.
Type: 'help' for help with commands
'quit' to quit
virsh #
[root@compute ~]# virsh -c qemu+tcp://controller/system
Welcome to virsh, the virtualization interactive terminal.
Type: 'help' for help with commands
'quit' to quit
virsh # 3、修改nova.conf配置文件，使其允许迁移。
[root@controller ~]# vi /etc/nova/nova.conf
[libvirt]
inject_key = True
live_migration_flag = VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_TUNNELLED
block_migration_flag = VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_NON_SHARED_INC
live_migration_downtime = 500
live_migration_downtime_steps = 10
live_migration_downtime_delay = 75
[root@compute ~]# vi /etc/nova/nova.conf
[libvirt]
inject_key = True
live_migration_flag = VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_TUNNELLED
block_migration_flag = VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_NON_SHARED_INC
live_migration_downtime = 500
live_migration_downtime_steps = 10
live_migration_downtime_delay = 75
重启两个节点nova的所有服务:
在[libvirtd]选项中：一定要把
block_migration_flag==VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_NON_SHARED_INC
中的 VIR_MIGRATE_TUNNELLED删除。
否则报错：
[root@controller nova]# tail -f nova-compute.log
2019-05-15 16:24:50.692 4833 ERROR nova.virt.libvirt.driver [req-254721d5-5269-4c08-8d06-92ec37f6a306 c86d50b40c7b4b858952bdd24f11f9f9 bd8bb97a45cb4e6781c7f51952d16ffa - - -] [instance: 02e0808d-dae1-42e9-ae91-0e0a00f0f524] Live Migration failure: Operation not supported: Selecting disks to migrate is not implemented for tunnelled migration
2019-05-15 16:24:50.752 4833 ERROR nova.virt.libvirt.driver [req-254721d5-5269-4c08-8d06-92ec37f6a306 c86d50b40c7b4b858952bdd24f11f9f9 bd8bb97a45cb4e6781c7f51952d16ffa - - -] [instance: 02e0808d-dae1-42e9-ae91-0e0a00f0f524] Migration operation has aborted
4、开始迁移：
[root@controller ~]# nova show test
+--------------------------------------+----------------------------------------------------------+
| Property | Value |
+--------------------------------------+----------------------------------------------------------+
| OS-DCF:diskConfig | AUTO |
| OS-EXT-AZ:availability_zone | nova |
| OS-EXT-SRV-ATTR:host | controller |
| OS-EXT-SRV-ATTR:hostname | test |
| OS-EXT-SRV-ATTR:hypervisor_hostname | controller |
| OS-EXT-SRV-ATTR:instance_name | instance-0000001c |
[root@controller ~]# nova live-migration 02e0808d-dae1-42e9-ae91-0e0a00f0f524 compute
[root@controller ~]# nova show test
+--------------------------------------+----------------------------------------------------------+
| Property | Value |
+--------------------------------------+----------------------------------------------------------+
| OS-DCF:diskConfig | AUTO |
| OS-EXT-AZ:availability_zone | nova |
| OS-EXT-SRV-ATTR:host | compute |
| OS-EXT-SRV-ATTR:hostname | test |
| OS-EXT-SRV-ATTR:hypervisor_hostname | compute |
| OS-EXT-SRV-ATTR:instance_name | instance-0000001c |
| OS-EXT-SRV-ATTR:kernel_id | |
| OS-EXT-SRV-ATTR:launch_index | 0 |
| OS-EXT-SRV-ATTR:ramdisk_id | |
| OS-EXT-SRV-ATTR:reservation_id | r-x3iil7sr |
| OS-EXT-SRV-ATTR:root_device_name | /dev/vda |
| OS-EXT-SRV-ATTR:user_data | - |
| OS-EXT-STS:power_state | 1 |
| OS-EXT-STS:task_state | - |
| OS-EXT-STS:vm_state | active |
| OS-SRV-USG:launched_at | 2019-05-15T18:58:31.000000 |
| OS-SRV-USG:terminated_at | - |
| accessIPv4 | |
| accessIPv6 | |
| config_drive | |
| created | 2019-05-15T18:58:19Z |
| description | test |
| flavor | m1.small (2) |
| hostId | 80b6b07ccdf4b55333337f3f2d1e1b687a97301580c11a27806327ac |
| host_status | UP |
| id | 02e0808d-dae1-42e9-ae91-0e0a00f0f524 |
| image | centos-7.2 (6c3b67e7-d765-4a8a-8d26-97a5fd52971e) |
| init-net1 network | 10.0.0.44 |
| key_name | - |
| locked | False |
| metadata | {} |
| name | test |
| os-extended-volumes:volumes_attached | [] |
| progress | 0 |
| security_groups | default |
| status | ACTIVE |
| tenant_id | bd8bb97a45cb4e6781c7f51952d16ffa |
| updated | 2019-05-15T19:48:35Z |
| user_id | c86d50b40c7b4b858952bdd24f11f9f9 |
+--------------------------------------+----------------------------------------------------------+到此热迁移完成；
二、冷迁移--冷迁移相对于热迁移要简单很多：
流程：将某计算节点的实例关机——>将/var/lib/nova/instances/下对应的实例使用scp命令配合 -r 参数传输到目的主机的/var/lib/nova/instances/——>使用chown命令修改文件所有者——>手动修改nova数据的的instances表的host与node字段——>将虚拟机开机——>冷迁移完成
[root@controller ~]# nova show test
+--------------------------------------+----------------------------------------------------------+
| Property | Value |
+--------------------------------------+----------------------------------------------------------+
| OS-DCF:diskConfig | AUTO |
| OS-EXT-AZ:availability_zone | nova |
| OS-EXT-SRV-ATTR:host | controller |
| OS-EXT-SRV-ATTR:hostname | test |
| OS-EXT-SRV-ATTR:hypervisor_hostname | controller |
看清楚这台实例实在controller上，
[root@controller ~]#nova stop ceph
[root@controller ~]# scp -r /var/lib/nova/instances/02e0808d-dae1-42e9-ae91-0e0a00f0f524/ compute:/var/lib/nova/instances/
[root@compute ~]#chown nova:nova /var/lib/nova/instances/02e0808d-dae1-42e9-ae91-0e0a00f0f524/
如果还需要其他权限使用 ：chmod +777 -R /var/lib/nova/instances/02e0808d-dae1-42e9-ae91-0e0a00f0f524/
修改数据库：
MariaDB [(none)]> update nova.instances set host='compute',node='compute' where uuid='02e0808d-dae1-42e9-ae91-0e0a00f0f524';
重启下服务：
[root@compute ~]#systemctl restart openstack-nova-compute
开启实例：
[root@controller ~]#nova start ceph
[root@controller ~]# nova show test
+--------------------------------------+----------------------------------------------------------+
| Property | Value |
+--------------------------------------+----------------------------------------------------------+
| OS-DCF:diskConfig | AUTO |
| OS-EXT-AZ:availability_zone | nova |
| OS-EXT-SRV-ATTR:host | compute |
| OS-EXT-SRV-ATTR:hostname | test |
| OS-EXT-SRV-ATTR:hypervisor_hostname | compute