如何删除 OpenStack Nova 僵尸实例

转自：http://www.vpsee.com/2011/11/how-to-delete-a-openstack-nova-zombie-instance/

前天强制重启一台 OpenStack Nova 控制结点以后发现虚拟机消失，但是 euca-describe-instances 命令显示 instances 仍然是 running 的状态，使用 euca-terminate-instances 终止命令仍然无效，暂时把这样的 instance 称作"僵尸实例（zombie instance）"：

# virsh list
Id Name State
----------------------------------
# euca-describe-instances
RESERVATION    r-bkl83j20    bangcloud    default
INSTANCE    i-0000001d    ami-    172.16.39.121    172.16.39.121    running    vpsee (vpseecloud, node00)                --10T12::12Z    nova    aki-    ami-
RESERVATION    r-j335q6ny    bangcloud    default
INSTANCE    i-0000001e    ami-    172.16.39.122    172.16.39.122    running    vpsee (vpseecloud, node00)                --10T12::27Z    nova    aki-    ami-00000000
# euca-terminate-instances i-0000001d
# euca-terminate-instances i-0000001e

和 删除 OpenStack Nova Volume 时遇到的 error_deleting 问题 这篇文章提到的解决办法一样，直接操作数据库来删除这2条僵尸实例的记录。登录 mysql，使用 nova 数据库，找出要删除 instance 的 id，然后删除：

# mysql -u root -p
Enter password:
mysql> use nova;
mysql> select * from instances;
mysql> delete from instances where id = '';

ERROR  (): Cannot delete or update a parent row: a foreign key constraint fails (`nova`.`virtual_interfaces`, CONSTRAINT `virtual_interfaces_ibfk_1` FOREIGN KEY (`instance_id`) REFERENCES `instances` (`id`))

MySQL 删除 id 为 29 的 instance 时触发外键限制错误，简单的办法是暂时关闭外键检查，等删除后再打开：

mysql> SET FOREIGN_KEY_CHECKS=;
Query OK,  rows affected (0.00 sec)

mysql> delete from instances where id = '';
Query OK,  row affected (0.04 sec)

mysql> delete from instances where id = '';
Query OK,  row affected (0.04 sec)

mysql> SET FOREIGN_KEY_CHECKS=;
Query OK,  rows affected (0.00 sec)

删除 instance 29 和 30后再用 euca-describe-instances 命令验证一下：

# euca-describe-instances