docs/userguide/opnfv_yardstick_tc025.rst

   1 .. This work is licensed under a Creative Commons Attribution 4.0 International
   2 .. License.
   3 .. http://creativecommons.org/licenses/by/4.0
   4 .. (c) OPNFV, Huawei Technologies Co.,Ltd and others.
   5
   6 *************************************
   7 Yardstick Test Case Description TC025
   8 *************************************
   9
  10 +-----------------------------------------------------------------------------+
  11 |OpenStack Controller Node abnormally shutdown High Availability              |
  12 |                                                                             |
  13 +--------------+--------------------------------------------------------------+
  14 |test case id  | OPNFV_YARDSTICK_TC025_HA: OpenStack Controller Node          |
  15 |              | abnormally shutdown                                          |
  16 |              |                                                              |
  17 +--------------+--------------------------------------------------------------+
  18 |test purpose  | This test case will verify the high availability of          |
  19 |              | controller node. When one of the controller node abnormally  |
  20 |              | shutdown, the service provided by it should be OK.           |
  21 |              |                                                              |
  22 +--------------+--------------------------------------------------------------+
  23 |test method   | This test case shutdowns a specified controller node with    |
  24 |              | some fault injection tools, then checks whether all services |
  25 |              | provided by the controller node are OK with some monitor     |
  26 |              | tools.                                                       |
  27 |              |                                                              |
  28 +--------------+--------------------------------------------------------------+
  29 |attackers     | In this test case, an attacker called "host-shutdown" 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 "host-shutdown" in       |
  33 |              | this test case.                                              |
  34 |              | 2) host: the name of a controller node being attacked.       |
  35 |              |                                                              |
  36 |              | e.g.                                                         |
  37 |              | -fault_type: "host-shutdown"                                 |
  38 |              | -host: node1                                                 |
  39 |              |                                                              |
  40 +--------------+--------------------------------------------------------------+
  41 |monitors      | In this test case, one kind of monitor are needed:           |
  42 |              | 1. the "openstack-cmd" monitor constantly request a specific |
  43 |              |    Openstack command, which needs two parameters             |
  44 |              | 1) monitor_type: which is used for finding the monitor class |
  45 |              | and related scritps. It should be always set to              |
  46 |              | "openstack-cmd" for this monitor.                            |
  47 |              | 2) command_name: which is the command name used for request  |
  48 |              |                                                              |
  49 |              | There are four instance of the "openstack-cmd" monitor:      |
  50 |              | monitor1:                                                    |
  51 |              | -monitor_type: "openstack-cmd"                               |
  52 |              | -api_name: "nova image-list"                                 |
  53 |              | monitor2:                                                    |
  54 |              | -monitor_type: "openstack-cmd"                               |
  55 |              | -api_name: "neutron router-list"                             |
  56 |              | monitor3:                                                    |
  57 |              | -monitor_type: "openstack-cmd"                               |
  58 |              | -api_name: "heat stack-list"                                 |
  59 |              | monitor4:                                                    |
  60 |              | -monitor_type: "openstack-cmd"                               |
  61 |              | -api_name: "cinder list"                                     |
  62 |              |                                                              |
  63 +--------------+--------------------------------------------------------------+
  64 |metrics       | In this test case, there is one metric:                      |
  65 |              | 1)service_outage_time: which indicates the maximum outage    |
  66 |              | time (seconds) of the specified Openstack command request.   |
  67 |              |                                                              |
  68 +--------------+--------------------------------------------------------------+
  69 |test tool     | Developed by the project. Please see folder:                 |
  70 |              | "yardstick/benchmark/scenarios/availability/ha_tools"        |
  71 |              |                                                              |
  72 +--------------+--------------------------------------------------------------+
  73 |references    | ETSI NFV REL001                                              |
  74 |              |                                                              |
  75 +--------------+--------------------------------------------------------------+
  76 |configuration | This test case needs two configuration files:                |
  77 |              | 1) test case file: opnfv_yardstick_tc019.yaml                |
  78 |              | -Attackers: see above "attackers" discription                |
  79 |              | -waiting_time: which is the time (seconds) from the process  |
  80 |              | being killed to stoping monitors the monitors                |
  81 |              | -Monitors: see above "monitors" discription                  |
  82 |              | -SLA: see above "metrics" discription                        |
  83 |              |                                                              |
  84 |              | 2)POD file: pod.yaml                                         |
  85 |              | The POD configuration should record on pod.yaml first.       |
  86 |              | the "host" item in this test case will use the node name in  |
  87 |              | the pod.yaml.                                                |
  88 |              |                                                              |
  89 +--------------+--------------------------------------------------------------+
  90 |test sequence | description and expected result                              |
  91 |              |                                                              |
  92 +--------------+--------------------------------------------------------------+
  93 |step 1        | start monitors:                                              |
  94 |              | each monitor will run with independently process             |
  95 |              |                                                              |
  96 |              | Result: The monitor info will be collected.                  |
  97 |              |                                                              |
  98 +--------------+--------------------------------------------------------------+
  99 |step 2        | do attacker: connect the host through SSH, and then execute  |
 100 |              | shutdown script on the host                                  |
 101 |              |                                                              |
 102 |              | Result: The host will be shutdown.                           |
 103 |              |                                                              |
 104 +--------------+--------------------------------------------------------------+
 105 |step 3        | stop monitors after a period of time specified by            |
 106 |              | "waiting_time"                                               |
 107 |              |                                                              |
 108 |              | Result: All monitor result will be aggregated.               |
 109 |              |                                                              |
 110 +--------------+--------------------------------------------------------------+
 111 |step 4        | verify the SLA                                               |
 112 |              |                                                              |
 113 |              | Result: The test case is passed or not.                      |
 114 |              |                                                              |
 115 +--------------+--------------------------------------------------------------+
 116 |post-action   | It is the action when the test cases exist. It restarts the  |
 117 |              | specified controller node if it is not restarted.            |
 118 |              |                                                              |
 119 +--------------+--------------------------------------------------------------+
 120 |test verdict  | Fails only if SLA is not passed, or if there is a test case  |
 121 |              | execution problem.                                           |
 122 |              |                                                              |
 123 +--------------+--------------------------------------------------------------+