作者：广东乐维软件有限公司 许远

1 概述**

OpenStack是一个开源的 云计算 管理平台项目，是一系列软件开源项目的组合。由 NASA (美国国家航空航天局)和 Rackspace 合作研发 并发起，以Apache许可证（ Apache软件基金会 发布的一个自由 软件许可证 ）授权。
OpenStack为私有云和公有云提供可扩展的弹性的 云计算服务 。 项目目标 是提供实施简单、可大规模扩展、丰富、标准统一的云计算管理平台。
对于 Openstack运维人员来说，需要掌握Openstack云平台系统资源的整体运行情况，包括域(AZ)的cpu/menmory等资源使用情况；本文将介绍使用zabbix结合openstack python-api对openstack系统资源进行监控 ， 详情如下

2 使用控制台获取域(AZ)监控信息

使用如下命令可以获取当前 openstack的所有可用域 。

nova availability-zone-list

+-----------------------+----------------------------------------+
| Name                  | Status                                 |
+-----------------------+----------------------------------------+
| internal              | available                              |
| |- computer03         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:07.000000 |
| |- computer04         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:08.000000 |
| |- computer05         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:13.000000 |
| |- computer06         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:05.000000 |
| |- computer07         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:05.000000 |
| |- computer08         |                                        |
| | |- nova-storage     | enabled :-) 2018-10-18T10:56:07.000000 |
| |- controler02        |                                        |
| | |- nova-conductor   | enabled XXX 2017-10-09T09:46:27.000000 |
| | |- nova-consoleauth | enabled XXX 2017-10-09T09:46:37.000000 |
| | |- nova-monitor     | enabled :-) 2018-10-18T10:56:11.000000 |
| | |- nova-scheduler   | enabled XXX 2017-10-09T09:46:37.000000 |
| | |- nova-cert        | enabled XXX 2017-10-09T09:46:37.000000 |
| |- controller01       |                                        |
| | |- nova-conductor   | enabled :-) 2018-10-18T10:56:11.000000 |
| | |- nova-consoleauth | enabled :-) 2018-10-18T10:56:04.000000 |
| | |- nova-monitor     | enabled :-) 2018-10-18T10:56:07.000000 |
| | |- nova-scheduler   | enabled :-) 2018-10-18T10:56:09.000000 |
| | |- nova-cert        | enabled :-) 2018-10-18T10:56:12.000000 |
| IMS                   | available                              |
| |- computer03         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:07.000000 |
| |- computer04         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:05.000000 |
| |- computer08         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:13.000000 |
| paas                  | available                              |
| |- computer05         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:07.000000 |
| |- computer06         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:13.000000 |
| |- computer07         |                                        |
| | |- nova-compute     | enabled :-) 2018-10-18T10:56:04.000000 |
+-----------------------+----------------------------------------+

由结果可以看出，该系统含有两个域： IMS 和 paas 。
但是该命令结果不太直观，不容易提取，我们也可以使用如下命令：

nova aggregate-list

+----+---------+-------------------+
| Id | Name    | Availability Zone |
+----+---------+-------------------+
| 1  | IMS+RCS | IMS               |
| 4  | paas    | paas              |
+----+---------+-------------------+

根据 aggregate-list ，我们可以查看域对应的计算节点 ：

nova aggregate-details paas
 +----+------+-------------------+------------------------------------------+--------------------------+
 | Id | Name | Availability Zone | Hosts | Metadata |
 +----+------+-------------------+------------------------------------------+--------------------------+
 | 4 | paas | paas | 'computer07', 'computer05', 'computer06' | 'availability_zone=paas' |
 +----+------+-------------------+------------------------------------------+--------------------------+

由结果可知， paas域，一共包含三个computer节点，对于每个节点的使用情况，我们一样可以通过命令获取 ：

nova hypervisor-show computer07
 +---------------------------+------------+
 | Property | Value |
 +---------------------------+------------+
 | free_disk_gb | 55 |
 | free_ram_mb | 126156 |
 | host_ip | 193.2.0.37 |
 | hypervisor_hostname | computer07 |
 | hypervisor_type | QEMU |
 | id | 9 |
 | local_gb | 445 |
 | local_gb_used | 390 |
 | memory_mb | 257740 |
 | memory_mb_used | 131584 |
 | npt_ept | ept |
 | pci_pools | - |
 | running_vms | 8 |
 | service_disabled_reason | - |
 | service_host | computer07 |
 | service_id | 27 |
 | state | up |
 | status | enabled |
 | vcpus | 48 |
 | vcpus_used | 32 |
 +---------------------------+------------+
 #截取部分信息

我们可以得知，该节点的 vcpu总数及其使用情况,memory_mb总数及其使用情况。
将每个节点的使用情况均获取后，经过计算，就可以得出域的资源整体使用情况。对于云平台的整体使用情况，我们也可以通过命令行获取：

nova hypervisor-stats
 +----------------------+---------+
 | Property | Value |
 +----------------------+---------+
 | count | 6 |
 | current_workload | 0 |
 | disk_available_least | 304 |
 | free_disk_gb | 518 |
 | free_ram_mb | 961736 |
 | local_gb | 2670 |
 | local_gb_used | 2152 |
 | memory_mb | 1546440 |
 | memory_mb_used | 584704 |
 | running_vms | 42 |
 | vcpus | 288 |
 | vcpus_used | 172 |
 +----------------------+---------+

上面我们讨论了使用命令行进行相关信息的获取，下面我们讨论使用 python-api进行相关信息的获取与计算。

3 使用openstack python-api 获取域监控信息

程序代码如下：

#!/usr/bin/python
# -*- coding: utf-8 -*-

#imports
import json
from optparse import OptionParser
from novaclient import client as noclient
from novaclient import utils

#登录及授权
keystone = {}
keystone['os_username']='admin'
keystone['os_password']='keystone'
keystone['os_auth_url']='http://lb-vip:5000/v2.0/'
keystone['os_tenant_name']='admin'
nova_client = noclient.Client(2, keystone['os_username'], keystone['os_password'], keystone['os_tenant_name'], keystone['os_auth_url'])

def main():
   #获取云平台整体信息并打印
   total_info = nova_client.hypervisor_stats.statistics()._info.copy()
   print "total_info_vcpus:", total_info["vcpus"]
   print "total_info_vcpus_used:", total_info["vcpus_used"]
   print "total_info_memory_mb:", total_info["memory_mb"]
   print "total_info_memory_mb_used:", total_info["memory_mb_used"]
   print "total_info_running_vms:", total_info["running_vms"]

   #获取域列表信息
   aggregates = nova_client.aggregates.list()

   for aggregate in aggregates:
       #初始化每个域的资源统计变量
       vcpus = 0
       vcpus_used = 0
       memory_mb = 0
       memory_mb_used = 0
       running_vms = 0

       #获取每个aggregate信息，并保存对应的hostscomputer节点列表
       aggregate_info = aggregate._info.copy()
       print aggregate_info["id"], aggregate_info["name"], aggregate_info["availability_zone"], aggregate_info["hosts"]
       aggregate_hosts = aggregate_info["hosts"]

       #循环计算节点，保存相关资源信息
       for aggregate_host in aggregate_hosts:
           hypervisor_info = utils.find_resource(nova_client.hypervisors, aggregate_host)._info
           vcpus = vcpus + hypervisor_info["vcpus"]
           vcpus_used = vcpus_used + hypervisor_info["vcpus_used"]
           memory_mb = memory_mb + hypervisor_info["memory_mb"]
           memory_mb_used = memory_mb_used + hypervisor_info["memory_mb_used"]
           running_vms = running_vms + hypervisor_info["running_vms"]
       
       #打印域资源信息
       print "vcpus:", vcpus
       print "vcpus_used:", vcpus_used
       print "memory_mb:", memory_mb
       print "memory_mb_used:", memory_mb_used
       print "running_vms:", running_vms

if __name__ == "__main__":
   main()

执行该程序后，可以获取各个域节点的信息：

total_info_vcpus: 288
total_info_vcpus_used: 172
total_info_memory_mb: 1546440
total_info_memory_mb_used: 584704
total_info_running_vms: 42

1 IMS+RCS IMS [u'computer04', u'computer03', u'computer08']
vcpus: 144
vcpus_used: 84
memory_mb: 773220
memory_mb_used: 243200
running_vms: 20

4 paas paas [u'computer07', u'computer05', u'computer06']
vcpus: 144
vcpus_used: 88
memory_mb: 773220
memory_mb_used: 341504
running_vms: 22

经过适当的计算，我们就可以获取各个域分配及使用比例等信息。
上面我们就使用 python-api 打印出了所有需要的信息，但是对于监控来说，我们需要提取的是各个监控项的信息，这样才能方便的搜索和做图表展示；下面我们讨论结合zabbix进行相关信息的监控。

4 结合zabbix获取域相关监控信息

4.1 获取可用域信息列表**

上面我们已经获取了所有的可用域信息，但对于 zabbix来说，我们还需要返回固定格式的数据，供zabbix进行解析：

#!/usr/bin/python
# -*- coding: utf-8 -*-

#imports
import json
from optparse import OptionParser
from novaclient import client as noclient
from novaclient import utils

#登录及授权
keystone = {}
keystone['os_username']='admin'
keystone['os_password']='keystone'
keystone['os_auth_url']='http://lb-vip:5000/v2.0/'
keystone['os_tenant_name']='admin'
nova_client = noclient.Client(2, keystone['os_username'], keystone['os_password'], keystone['os_tenant_name'], keystone['os_auth_url'])

def main():
   r = {"data":[]}
   aggregates = nova_client.aggregates.list()

   for aggregate in aggregates:
       aggregate_info = aggregate._info.copy()
       r['data'].append( {"{#NAME}":aggregate_info["name"]} )
   print(json.dumps(r, indent=2, sort_keys=True, encoding="utf-8"))

if __name__ == "__main__":
   main()

执行后，返回的结果如下 ,可以通过供zabbix自动发现模版使用 ：

{
  "data": [
    {
      "{#AVAILABLE_ZONE}": "IMS",
      "{#NAME}": "IMS+RCS"
    },
    {
      "{#AVAILABLE_ZONE}": "paas",
      "{#NAME}": "paas"
    }
  ]
}

4.2 zabbix，Item及自动发现LLD配置**

云平台整体信息， Item设置如下：
openstack.total[vcpus]
openstack.total[vcpus_used]
openstack.total[memory_mb]
openstack.total[memory_mb_used]
openstack.total[running_vms]
获取 {#NAME}后就可以根据监控项获取对应的监控内容了，zabbix自动发现自动发现key设置如下：

discovery下，设置监控Item,key设置如下：

其中 {#NAME}为第一步的脚本中返回的可用域对应的aggregate相关信息。

4.3 zabbix-agent配置与对应的脚本**

对应的 zabbix-agent.conf配置如下：

# /etc/zabbix/zabbix_agentd.d/userparameter_openstack-system.conf

UserParameter=openstack.system.discovery,python /etc/zabbix/zabbix_agentd.d/openstack-system.py --item discovery

UserParameter=openstack.total[*],python /etc/zabbix/zabbix_agentd.d/openstack-system.py --item total --moniter $1

UserParameter=openstack.zone[*],python /etc/zabbix/zabbix_agentd.d/openstack-system.py --item $1 --aggregate $2

对应的 zabbix自动发现的监控脚本如下：

# /etc/zabbix/zabbix_agentd.d/openstack-system.py

#!/usr/bin/python
# -*- coding: utf-8 -*-

#imports
import json
from optparse import OptionParser
from novaclient import client as noclient
from novaclient import utils

#登录及授权
keystone = {}
keystone['os_username']='admin'
keystone['os_password']='keystone'
keystone['os_auth_url']='http://lb-vip:5000/v2.0/'
keystone['os_tenant_name']='admin'
nova_client = noclient.Client(2, keystone['os_username'], keystone['os_password'], keystone['os_tenant_name'], keystone['os_auth_url'])

def main():
    options = parse_args()
    if options.item=="discovery":
        zone_list()
    elif options.item=="total":
        total_moniter(options)
    else:
        zone_moniter(options)

#判断入参合法性
def parse_args():
    parser = OptionParser()
    valid_item = ["discovery", "total", "hosts", "vcpus", "vcpus_used", "memory_mb", "memory_mb_used", "running_vms"]
    parser.add_option("", "--item", dest="item", help="", action="store", type="string", default=None)
    parser.add_option("", "--moniter", dest="moniter", help="", action="store", type="string", default=None)
    parser.add_option("", "--aggregate", dest="aggregate", help="", action="store", type="string", default=None)
    (options, args) = parser.parse_args()
    if options.item not in valid_item:
        parser.error("Item has to be one of: "+", ".join(valid_item))
    return options

#获取可用域列表
def zone_list():
    r = {"data":[]}
    aggregates = nova_client.aggregates.list()

    for aggregate in aggregates:
        aggregate_info = aggregate._info.copy()
        r['data'].append( {"{#NAME}":aggregate_info["name"], "{#AVAILABLE_ZONE}":aggregate_info["availability_zone"]} )
    print(json.dumps(r, indent=2, sort_keys=True, encoding="utf-8"))

#获取云平台整体监控信息
def total_moniter(options):
    total_info = nova_client.hypervisor_stats.statistics()._info.copy()
    print (total_info[options.moniter])

#获取可用域对应的监控信息
def zone_moniter(options):
    aggregate = utils.find_resource(nova_client.aggregates, options.aggregate)
    aggregate_info = aggregate._info.copy()
    aggregate_hosts = aggregate_info["hosts"]
    if options.item=="hosts":
        print (aggregate_hosts)
    else:
        monitor_data = 0
        for aggregate_host in aggregate_hosts:
            hypervisor_info = utils.find_resource(nova_client.hypervisors, aggregate_host)._info
            monitor_data = monitor_data + hypervisor_info[options.item]
        print (monitor_data)

if __name__ == "__main__":
    main()

在 zabbix上进行对应的配置后重启，将模版应用于主机，此时应当监控获取所有的可用域，并监控对应的信息。

1 参考资料**

1. Openstack 中的zone ,aggregates和host及其应用,https://blog.csdn.net/ztejiagn/article/details/8948688
2. nova 命令汇总四 ——计算相关命令,http://blog.51cto.com/13788458/2129157
3. The novaclient Python API,https://docs.openstack.org/python-novaclient/latest/reference/api/index.html
4. GitHub - larsks/openstack-api-samples,https://github.com/larsks/openstack-api-samples
   以上就是这一期的Zabbix技术分享内容。