1 .. This work is licensed under a Creative Commons Attribution 4.0 International
3 .. http://creativecommons.org/licenses/by/4.0
4 .. (c) OPNFV, Yin Kanglin and others.
5 .. 14_ykl@tongji.edu.cn
7 *************************************
8 Yardstick Test Case Description TC051
9 *************************************
11 +-----------------------------------------------------------------------------+
12 |OpenStack Controller Node CPU Overload High Availability |
14 +--------------+--------------------------------------------------------------+
15 |test case id | OPNFV_YARDSTICK_TC051: OpenStack Controller Node CPU |
16 | | Overload High Availability |
17 +--------------+--------------------------------------------------------------+
18 |test purpose | This test case will verify the high availability of control |
19 | | node. When the CPU usage of a specified controller node is |
20 | | stressed to 100%, which breaks down the Openstack services |
21 | | on this node. These Openstack service should able to be |
22 | | accessed by other controller nodes, and the services on |
23 | | failed controller node should be isolated. |
24 +--------------+--------------------------------------------------------------+
25 |test method | This test case stresses the CPU uasge of a specified control |
26 | | node to 100%, then checks whether all services provided by |
27 | | the environment are OK with some monitor tools. |
28 +--------------+--------------------------------------------------------------+
29 |attackers | In this test case, an attacker called "stress-cpu" is |
30 | | needed. This attacker includes two parameters: |
31 | | 1) fault_type: which is used for finding the attacker's |
32 | | scripts. It should be always set to "stress-cpu" in |
34 | | 2) host: which is the name of a control node being attacked. |
36 | | -fault_type: "stress-cpu" |
38 +--------------+--------------------------------------------------------------+
39 |monitors | In this test case, the monitor named "openstack-cmd" is |
40 | | needed. The monitor needs needs two parameters: |
41 | | 1) monitor_type: which is used for finding the monitor class |
42 | | and related scritps. It should be always set to |
43 | | "openstack-cmd" for this monitor. |
44 | | 2) command_name: which is the command name used for request |
46 | | There are four instance of the "openstack-cmd" monitor: |
48 | | -monitor_type: "openstack-cmd" |
49 | | -command_name: "nova image-list" |
51 | | -monitor_type: "openstack-cmd" |
52 | | -command_name: "neutron router-list" |
54 | | -monitor_type: "openstack-cmd" |
55 | | -command_name: "heat stack-list" |
57 | | -monitor_type: "openstack-cmd" |
58 | | -command_name: "cinder list" |
59 +--------------+--------------------------------------------------------------+
60 |metrics | In this test case, there is one metric: |
61 | | 1)service_outage_time: which indicates the maximum outage |
62 | | time (seconds) of the specified Openstack command request. |
63 +--------------+--------------------------------------------------------------+
64 |test tool | Developed by the project. Please see folder: |
65 | | "yardstick/benchmark/scenarios/availability/ha_tools" |
67 +--------------+--------------------------------------------------------------+
68 |references | ETSI NFV REL001 |
70 +--------------+--------------------------------------------------------------+
71 |configuration | This test case needs two configuration files: |
72 | | 1) test case file: opnfv_yardstick_tc051.yaml |
73 | | -Attackers: see above "attackers" discription |
74 | | -waiting_time: which is the time (seconds) from the process |
75 | | being killed to stoping monitors the monitors |
76 | | -Monitors: see above "monitors" discription |
77 | | -SLA: see above "metrics" discription |
79 | | 2)POD file: pod.yaml |
80 | | The POD configuration should record on pod.yaml first. |
81 | | the "host" item in this test case will use the node name in |
84 +--------------+--------------------------------------------------------------+
85 |test sequence | description and expected result |
87 +--------------+--------------------------------------------------------------+
88 |step 1 | start monitors: |
89 | | each monitor will run with independently process |
91 | | Result: The monitor info will be collected. |
93 +--------------+--------------------------------------------------------------+
94 |step 2 | do attacker: connect the host through SSH, and then execute |
95 | | the stress cpu script on the host. |
97 | | Result: The CPU usage of the host will be stressed to 100%. |
99 +--------------+--------------------------------------------------------------+
100 |step 3 | stop monitors after a period of time specified by |
103 | | Result: The monitor info will be aggregated. |
105 +--------------+--------------------------------------------------------------+
106 |step 4 | verify the SLA |
108 | | Result: The test case is passed or not. |
110 +--------------+--------------------------------------------------------------+
111 |post-action | It is the action when the test cases exist. It kills the |
112 | | process that stresses the CPU usage. |
113 +--------------+--------------------------------------------------------------+
114 |test verdict | Fails only if SLA is not passed, or if there is a test case |
115 | | execution problem. |
117 +--------------+--------------------------------------------------------------+