Imporve monitor_process pass criteria
[yardstick.git] / yardstick / benchmark / scenarios / availability / serviceha.py
index 3e03e1d..2e82971 100755 (executable)
@@ -6,13 +6,12 @@
 # which accompanies this distribution, and is available at
 # http://www.apache.org/licenses/LICENSE-2.0
 ##############################################################################
-import pkg_resources
+from __future__ import print_function
+from __future__ import absolute_import
 import logging
-import time
-import yaml
-import yardstick.ssh as ssh
 from yardstick.benchmark.scenarios import base
-from yardstick.benchmark.scenarios.availability import monitor
+from yardstick.benchmark.scenarios.availability.monitor import basemonitor
+from yardstick.benchmark.scenarios.availability.attacker import baseattacker
 
 LOG = logging.getLogger(__name__)
 
@@ -22,86 +21,36 @@ class ServiceHA(base.Scenario):
     """
     __scenario_type__ = "ServiceHA"
 
-    HA_CONF = "ha_tools/ha_conf.yaml"
-
     def __init__(self, scenario_cfg, context_cfg):
+        LOG.debug(
+            "scenario_cfg:%s context_cfg:%s",
+            scenario_cfg, context_cfg)
         self.scenario_cfg = scenario_cfg
         self.context_cfg = context_cfg
-        self.service_name = scenario_cfg["options"]["component"]
-        self.fault_type = scenario_cfg["options"]["fault_type"]
-        self.fault_time = scenario_cfg["options"].get("fault_time", 0)
-        self.fault_cfg = None
         self.setup_done = False
-        self.need_teardown = False
+        self.data = {}
 
     def setup(self):
-        '''scenario setup'''
-        self.ha_conf_file = pkg_resources.resource_filename(
-            "yardstick.benchmark.scenarios.availability",
-            ServiceHA.HA_CONF)
-        ha_cfg = []
-        with open(self.ha_conf_file) as stream:
-                ha_cfg = yaml.load(stream)
-        LOG.debug("ha_cfg content:%s" % ha_cfg)
-
-        # check the ha_conf contains the service defined in test cases yaml
-        service_cfg = ha_cfg.get(self.service_name, None)
-        if not service_cfg:
-            LOG.error(
-                "The component %s can not be supported!" % self.service_name)
+        """scenario setup"""
+        nodes = self.context_cfg.get("nodes", None)
+        if nodes is None:
+            LOG.error("the nodes info is none")
             return
 
-        for fault in service_cfg:
-            if fault["type"] == self.fault_type:
-                self.fault_cfg = fault
-                break
-        if not self.fault_cfg:
-            LOG.error(
-                "The fualt_type %s can not be supproted!" % self.fault_type)
-            return
-        LOG.debug("the fault_cfg :%s" % self.fault_cfg)
-
-        self.fault_script = pkg_resources.resource_filename(
-            "yardstick.benchmark.scenarios.availability",
-            self.fault_cfg["inject_script"])
-        self.recovery_script = pkg_resources.resource_filename(
-            "yardstick.benchmark.scenarios.availability",
-            self.fault_cfg["recovery_script"])
-        self.check_script = pkg_resources.resource_filename(
-            "yardstick.benchmark.scenarios.availability",
-            self.fault_cfg["check_script"])
-
-        host = self.context_cfg.get("host", None)
-        ip = host.get("ip", None)
-        user = host.get("user", "root")
-        key_filename = host.get("key_filename", "~/.ssh/id_rsa")
-        LOG.info("The host: %s  the service: %s" % (ip, self.service_name))
-        LOG.debug("The params, host:%s  fault_cfg:%s" % (host, self.fault_cfg))
+        self.attackers = []
+        attacker_cfgs = self.scenario_cfg["options"]["attackers"]
+        for attacker_cfg in attacker_cfgs:
+            attacker_cls = baseattacker.BaseAttacker.get_attacker_cls(
+                attacker_cfg)
+            attacker_ins = attacker_cls(attacker_cfg, nodes)
+            attacker_ins.setup()
+            self.attackers.append(attacker_ins)
+            self.data = dict(self.data.items() + attacker_ins.data.items())
 
-        LOG.debug(
-            "ssh connection ip:%s, user:%s, key_file:%s",
-            ip, user, key_filename)
-        self.connection = ssh.SSH(user, ip, key_filename=key_filename)
-        self.connection.wait(timeout=600)
-        LOG.debug("ssh host success!")
-
-        # check the host envrioment
-        exit_status, stdout, stderr = self.connection.execute(
-            "/bin/sh -s {0}".format(self.service_name),
-            stdin=open(self.check_script, "r"))
-        LOG.info(
-            "the exit_status:%s stdout:%s stderr:%s" %
-            (exit_status, stdout, stderr))
-        if exit_status:
-            raise RuntimeError(stderr)
-
-        if stdout and "running" in stdout:
-            LOG.info("check the envrioment success!")
-        else:
-            LOG.error(
-                "the host envrioment is error, stdout:%s, stderr:%s" %
-                (stdout, stderr))
-            return
+        monitor_cfgs = self.scenario_cfg["options"]["monitors"]
+
+        self.monitorMgr = basemonitor.MonitorMgr(self.data)
+        self.monitorMgr.init_monitors(monitor_cfgs, nodes)
 
         self.setup_done = True
 
@@ -111,83 +60,76 @@ class ServiceHA(base.Scenario):
             LOG.error("The setup not finished!")
             return
 
-        monitorInstance = monitor.Monitor()
-        monitorInstance.setup(self.fault_cfg)
-        monitorInstance.start()
-        LOG.info("monitor start!")
-
-        LOG.info("Inject fault!")
-        exit_status, stdout, stderr = self.connection.execute(
-            "/bin/sh -s {0}".format(self.service_name),
-            stdin=open(self.fault_script, "r"))
-
-        if exit_status != 0:
-            monitorInstance.stop()
-            raise RuntimeError(stderr)
-
-        self.need_teardown = True
-        time.sleep(self.fault_time)
-
-        monitorInstance.stop()
-        LOG.info("monitor stop!")
-
-        ret = monitorInstance.get_result()
-        LOG.info("The monitor result:%s" % ret)
-        outage_time = ret.get("outage_time")
-        result["outage_time"] = outage_time
-        LOG.info("the result:%s" % result)
-
-        if "sla" in self.scenario_cfg:
-            sla_outage_time = int(self.scenario_cfg["sla"]["outage_time"])
-            assert outage_time <= sla_outage_time, "outage_time %f > sla:outage_time(%f)" % \
-                (outage_time, sla_outage_time)
+        self.monitorMgr.start_monitors()
+        LOG.info("HA monitor start!")
+
+        for attacker in self.attackers:
+            attacker.inject_fault()
+
+        self.monitorMgr.wait_monitors()
+        LOG.info("HA monitor stop!")
+
+        sla_pass = self.monitorMgr.verify_SLA()
+        for k, v in self.data.items():
+            if self.data[k] == 0:
+                result['sla_pass'] = 0
+                LOG.info("The service process not found in the host envrioment, \
+the HA test case NOT pass")
+                return
+        if sla_pass:
+            result['sla_pass'] = 1
+            LOG.info("The HA test case PASS the SLA")
+        else:
+            result['sla_pass'] = 0
+        assert sla_pass is True, "The HA test case NOT pass the SLA"
 
         return
 
     def teardown(self):
-        '''scenario teardown'''
-        LOG.info("recory the everiment!")
-
-        if self.need_teardown:
-            exit_status, stdout, stderr = self.connection.execute(
-                "/bin/sh -s {0} ".format(self.service_name),
-                stdin=open(self.recovery_script, "r"))
-
-            if exit_status:
-                raise RuntimeError(stderr)
-            else:
-                self.need_teardown = False
-
-"""
-def _test():
-    '''internal test function'''
+        """scenario teardown"""
+        for attacker in self.attackers:
+            attacker.recover()
+
+
+def _test():    # pragma: no cover
+    """internal test function"""
     host = {
         "ip": "10.20.0.5",
         "user": "root",
         "key_filename": "/root/.ssh/id_rsa"
     }
-    ctx = {"host": host}
-
-    logger = logging.getLogger("yardstick")
-    logger.setLevel(logging.DEBUG)
+    ctx = {"nodes": {"node1": host}}
+    attacker_cfg = {
+        "fault_type": "kill-process",
+        "process_name": "nova-api",
+        "host": "node1"
+    }
+    attacker_cfgs = []
+    attacker_cfgs.append(attacker_cfg)
+    monitor_cfg = {
+        "monitor_cmd": "nova image-list"
+    }
+    monitor_cfgs = []
+    monitor_cfgs.append(monitor_cfg)
 
     options = {
-        "component": "nova-api",
-        "fault_type": "stop-service"
+        "attackers": attacker_cfgs,
+        "wait_time": 10,
+        "monitors": monitor_cfgs
     }
     sla = {"outage_time": 5}
     args = {"options": options, "sla": sla}
 
-    print "create instance"
+    print("create instance")
     terstInstance = ServiceHA(args, ctx)
 
     terstInstance.setup()
     result = {}
     terstInstance.run(result)
-    print result
+    print(result)
 
     terstInstance.teardown()
 
-if __name__ == '__main__':
+
+if __name__ == '__main__':    # pragma: no cover
     _test()
-"""