docs/testing/user/userguide/opnfv_yardstick_tc090.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, Yin Kanglin and others.
   5 .. 14_ykl@tongji.edu.cn
   6
   7 *************************************
   8 Yardstick Test Case Description TC090
   9 *************************************
  10
  11 +-----------------------------------------------------------------------------+
  12 |Control Node OpenStack Service High Availability - Database Instances        |
  13 |                                                                             |
  14 +--------------+--------------------------------------------------------------+
  15 |test case id  | OPNFV_YARDSTICK_TC090: Control node OpenStack service down - |
  16 |              | database instances                                           |
  17 +--------------+--------------------------------------------------------------+
  18 |test purpose  | This test case will verify the high availability of the      |
  19 |              | data base instances used by OpenStack (mysql) on control     |
  20 |              | node.                                                        |
  21 |              |                                                              |
  22 +--------------+--------------------------------------------------------------+
  23 |test method   | This test case kills the processes of database service on a  |
  24 |              | selected control node, then checks whether the request of    |
  25 |              | the related OpenStack command is OK and the killed processes |
  26 |              | are recovered.                                               |
  27 |              |                                                              |
  28 +--------------+--------------------------------------------------------------+
  29 |attackers     | In this test case, an attacker called "kill-process" is      |
  30 |              | needed. This attacker includes three parameters:             |
  31 |              | 1) fault_type: which is used for finding the attacker's      |
  32 |              | scripts. It should be always set to "kill-process" in this   |
  33 |              | test case.                                                   |
  34 |              | 2) process_name: which is the process name of the specified  |
  35 |              | OpenStack service. If there are multiple processes use the   |
  36 |              | same name on the host, all of them are killed by this        |
  37 |              | attacker.                                                    |
  38 |              | In this case. This parameter should always set to the name   |
  39 |              | of the database service of OpenStack.                        |
  40 |              | 3) host: which is the name of a control node being attacked. |
  41 |              |                                                              |
  42 |              | e.g.                                                         |
  43 |              | -fault_type: "kill-process"                                  |
  44 |              | -process_name: "mysql"                                       |
  45 |              | -host: node1                                                 |
  46 |              |                                                              |
  47 +--------------+--------------------------------------------------------------+
  48 |monitors      | In this test case, two kinds of monitor are needed:          |
  49 |              | 1. the "openstack-cmd" monitor constantly request a specific |
  50 |              | Openstack command, which needs two parameters:               |
  51 |              | 1) monitor_type: which is used for finding the monitor class |
  52 |              | and related scritps. It should be always set to              |
  53 |              | "openstack-cmd" for this monitor.                            |
  54 |              | 2) command_name: which is the command name used for request. |
  55 |              | In this case, the command name should be neutron related     |
  56 |              | commands.                                                    |
  57 |              |                                                              |
  58 |              | 2. the "process" monitor check whether a process is running  |
  59 |              | on a specific node, which needs three parameters:            |
  60 |              | 1) monitor_type: which used for finding the monitor class and|
  61 |              | related scripts. It should be always set to "process"        |
  62 |              | for this monitor.                                            |
  63 |              | 2) process_name: which is the process name for monitor       |
  64 |              | 3) host: which is the name of the node running the process   |
  65 |              |                                                              |
  66 |              | The examples of monitors show as follows, there are four     |
  67 |              | instance of the "openstack-cmd" monitor, in order to check   |
  68 |              | the database connection of different OpenStack components.   |
  69 |              |                                                              |
  70 |              | monitor1:                                                    |
  71 |              | -monitor_type: "openstack-cmd"                               |
  72 |              | -api_name: "openstack image list"                            |
  73 |              | monitor2:                                                    |
  74 |              | -monitor_type: "openstack-cmd"                               |
  75 |              | -api_name: "openstack router list"                           |
  76 |              | monitor3:                                                    |
  77 |              | -monitor_type: "openstack-cmd"                               |
  78 |              | -api_name: "openstack stack list"                            |
  79 |              | monitor4:                                                    |
  80 |              | -monitor_type: "openstack-cmd"                               |
  81 |              | -api_name: "openstack volume list"                           |
  82 |              | monitor5:                                                    |
  83 |              | -monitor_type: "process"                                     |
  84 |              | -process_name: "mysql"                                       |
  85 |              | -host: node1                                                 |
  86 |              |                                                              |
  87 +--------------+--------------------------------------------------------------+
  88 |metrics       | In this test case, there are two metrics:                    |
  89 |              | 1)service_outage_time: which indicates the maximum outage    |
  90 |              | time (seconds) of the specified OpenStack command request.   |
  91 |              | 2)process_recover_time: which indicates the maximum time     |
  92 |              | (seconds) from the process being killed to recovered         |
  93 |              |                                                              |
  94 +--------------+--------------------------------------------------------------+
  95 |test tool     | Developed by the project. Please see folder:                 |
  96 |              | "yardstick/benchmark/scenarios/availability/ha_tools"        |
  97 |              |                                                              |
  98 +--------------+--------------------------------------------------------------+
  99 |references    | ETSI NFV REL001                                              |
 100 |              |                                                              |
 101 +--------------+--------------------------------------------------------------+
 102 |configuration | This test case needs two configuration files:                |
 103 |              | 1) test case file: opnfv_yardstick_tc090.yaml                |
 104 |              | -Attackers: see above "attackers" description                |
 105 |              | -waiting_time: which is the time (seconds) from the process  |
 106 |              | being killed to stopping monitors the monitors               |
 107 |              | -Monitors: see above "monitors" description                  |
 108 |              | -SLA: see above "metrics" description                        |
 109 |              |                                                              |
 110 |              | 2)POD file: pod.yaml                                         |
 111 |              | The POD configuration should record on pod.yaml first.       |
 112 |              | the "host" item in this test case will use the node name in  |
 113 |              | the pod.yaml.                                                |
 114 |              |                                                              |
 115 +--------------+--------------------------------------------------------------+
 116 |test sequence | description and expected result                              |
 117 |              |                                                              |
 118 +--------------+--------------------------------------------------------------+
 119 |step 1        | start monitors:                                              |
 120 |              | each monitor will run with independently process             |
 121 |              |                                                              |
 122 |              | Result: The monitor info will be collected.                  |
 123 |              |                                                              |
 124 +--------------+--------------------------------------------------------------+
 125 |step 2        | do attacker: connect the host through SSH, and then execute  |
 126 |              | the kill process script with param value specified by        |
 127 |              | "process_name"                                               |
 128 |              |                                                              |
 129 |              | Result: Process will be killed.                              |
 130 |              |                                                              |
 131 +--------------+--------------------------------------------------------------+
 132 |step 3        | stop monitors after a period of time specified by            |
 133 |              | "waiting_time"                                               |
 134 |              |                                                              |
 135 |              | Result: The monitor info will be aggregated.                 |
 136 |              |                                                              |
 137 +--------------+--------------------------------------------------------------+
 138 |step 4        | verify the SLA                                               |
 139 |              |                                                              |
 140 |              | Result: The test case is passed or not.                      |
 141 |              |                                                              |
 142 +--------------+--------------------------------------------------------------+
 143 |post-action   | It is the action when the test cases exist. It will check the|
 144 |              | status of the specified process on the host, and restart the |
 145 |              | process if it is not running for next test cases             |
 146 |              |                                                              |
 147 +--------------+--------------------------------------------------------------+
 148 |test verdict  | Fails only if SLA is not passed, or if there is a test case  |
 149 |              | execution problem.                                           |
 150 |              |                                                              |
 151 +--------------+--------------------------------------------------------------+