k8s openstack ccm遇到的一些问题

1. node名 虚拟机名过长则无法加入集群

2. node网卡不一定会选到eth0作为管理网卡，应该是基于挂载序列选取，而非基于neutron接口顺序选第一个

(py3env) [root@ccn01 ~]# openstack port list  --server 41674d92-875b-42bc-9373-7dd8bb8afde2
+--------------------------------------+----------------------------------------------------------+-------------------+------------------------------------------------------------------------------+--------+
| ID                                   | Name                                                     | MAC Address       | Fixed IP Addresses                                                           | Status |
+--------------------------------------+----------------------------------------------------------+-------------------+------------------------------------------------------------------------------+--------+
| 7d3225bc-f9e9-46a2-ac14-d9c3cce207a8 | inner-prod-common-c6-4xl-asg-ofg-7pp-wa2-pubnet_port-i52 | fa:16:3e:c3:85:6d | ip_address='10.120.34.95', subnet_id='a69151b6-031e-40d7-83d6-67d49d4b6db6'  | ACTIVE |
| afb3fa7e-c9df-4cee-9cff-f58e91080bfe | pro-k8s-master-1-eth0                                    | fa:16:3e:62:63:a2 | ip_address='10.120.33.146', subnet_id='a69151b6-031e-40d7-83d6-67d49d4b6db6' | ACTIVE |
+--------------------------------------+----------------------------------------------------------+-------------------+------------------------------------------------------------------------------+--------+

# 比如这个接口 eth0 不在第一个，那么ccm将第一个返回poet作为主网卡，就会导致控制面面的ip是错的

3.跨租户的虚拟机无法加入集群，会被踢掉