add notification_driver config in nova.conf
[doctor.git] / tests / run.sh
index bbe093a..292a7ce 100755 (executable)
@@ -1,4 +1,4 @@
-#!/bin/bash -ex
+#!/bin/bash -e
 ##############################################################################
 # Copyright (c) 2016 NEC Corporation and others.
 #
@@ -8,6 +8,8 @@
 # http://www.apache.org/licenses/LICENSE-2.0
 ##############################################################################
 
+[[ "${CI_DEBUG:-true}" == [Tt]rue ]] && set -x
+
 IMAGE_URL=https://launchpad.net/cirros/trunk/0.3.0/+download/cirros-0.3.0-x86_64-disk.img
 IMAGE_NAME=cirros
 IMAGE_FILE="${IMAGE_NAME}.img"
@@ -17,55 +19,236 @@ VM_FLAVOR=m1.tiny
 ALARM_NAME=doctor_alarm1
 INSPECTOR_PORT=12345
 CONSUMER_PORT=12346
-
-SUPPORTED_INSTALLER_TYPES="apex local"
-INSTALLER_TYPE=${INSTALLER_TYPE:-apex}
+DOCTOR_USER=doctor
+DOCTOR_PW=doctor
+DOCTOR_PROJECT=doctor
+#TODO: change back to `_member_` when JIRA DOCTOR-55 is done
+DOCTOR_ROLE=admin
+
+SUPPORTED_INSTALLER_TYPES="apex fuel local"
+INSTALLER_TYPE=${INSTALLER_TYPE:-local}
 INSTALLER_IP=${INSTALLER_IP:-none}
-COMPUTE_HOST=${COMPUTE_HOST:-overcloud-novacompute-0}
-COMPUTE_IP=${COMPUTE_IP:-none}
-COMPUTE_USER=${COMPUTE_USER:-heat-admin}
+
+SUPPORTED_INSPECTOR_TYPES="sample congress"
+INSPECTOR_TYPE=${INSPECTOR_TYPE:-sample}
+
 ssh_opts="-o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no"
+as_doctor_user="--os-username $DOCTOR_USER --os-password $DOCTOR_PW
+                --os-tenant-name $DOCTOR_PROJECT"
 
 if [[ ! "$SUPPORTED_INSTALLER_TYPES" =~ "$INSTALLER_TYPE" ]] ; then
     echo "ERROR: INSTALLER_TYPE=$INSTALLER_TYPE is not supported."
     exit 1
 fi
 
-prepare_compute_ssh() {
-    ssh_opts_cpu="$ssh_opts"
+if [[ ! "$SUPPORTED_INSPECTOR_TYPES" =~ "$INSPECTOR_TYPE" ]] ; then
+    echo "ERROR: INSPECTOR_TYPE=$INSPECTOR_TYPE is not supported."
+    exit 1
+fi
 
+get_installer_ip() {
     if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
         if [[ "$INSTALLER_IP" == "none" ]] ; then
             instack_mac=$(sudo virsh domiflist instack | awk '/default/{print $5}')
             INSTALLER_IP=$(/usr/sbin/arp -e | grep ${instack_mac} | awk '{print $1}')
         fi
+    elif [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        if [[ "$INSTALLER_IP" == "none" ]] ; then
+            instack_mac=$(sudo virsh domiflist fuel-opnfv | awk '/pxebr/{print $5}')
+            INSTALLER_IP=$(/usr/sbin/arp -e | grep ${instack_mac} | awk '{print $1}')
+        fi
+    fi
 
-        if [[ "$COMPUTE_IP" == "none" ]] ; then
-            COMPUTE_IP=$(sudo ssh $ssh_opts $INSTALLER_IP \
-                         "source stackrc; \
-                          nova show $COMPUTE_HOST \
-                          | awk '/ ctlplane network /{print \$5}'")
+    if [[ "$INSTALLER_TYPE" != "local" ]] ; then
+        if [[ -z "$INSTALLER_IP" ]] ; then
+            echo "ERROR: no installer ip"
+            exit 1
         fi
+    fi
+}
+
+prepare_ssh_to_cloud() {
+    ssh_opts_cpu="$ssh_opts"
 
-        # get ssh key from installer node
+    # get ssh key from installer node
+    if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
         sudo scp $ssh_opts root@"$INSTALLER_IP":/home/stack/.ssh/id_rsa instack_key
         sudo chown $(whoami):$(whoami) instack_key
         chmod 400 instack_key
         ssh_opts_cpu+=" -i instack_key"
+    elif [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        sshpass -p r00tme scp $ssh_opts root@${INSTALLER_IP}:.ssh/id_rsa instack_key
+        sudo chown $(whoami):$(whoami) instack_key
+        chmod 400 instack_key
+        ssh_opts_cpu+=" -i instack_key"
     elif [[ "$INSTALLER_TYPE" == "local" ]] ; then
-        if [[ "$COMPUTE_IP" == "none" ]] ; then
-            COMPUTE_IP=$(getent hosts "$COMPUTE_HOST" | awk '{ print $1 }')
-            if [[ -z "$COMPUTE_IP" ]]; then
-                echo "ERROR: Could not resolve $COMPUTE_HOST. Either manually set COMPUTE_IP or enable DNS resolution."
-                exit 1
-            fi
+        echo "INSTALLER_TYPE set to 'local'. Assuming SSH keys already exchanged with $COMPUTE_HOST"
+    fi
+}
+
+prepare_test_env() {
+    #TODO delete it when fuel support the configuration
+    if [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        echo "modify the configuration..."
+        cat > set_conf.sh << 'END_TXT'
+#!/bin/bash
+if [ -e /etc/ceilometer/event_pipeline.yaml ]; then
+    if ! grep -q '^ *- notifier://?topic=alarm.all$' /etc/ceilometer/event_pipeline.yaml; then
+        sed -i 's|- notifier://|- notifier://?topic=alarm.all|' /etc/ceilometer/event_pipeline.yaml
+        echo "modify the ceilometer config"
+        service ceilometer-agent-notification restart
+    fi
+else
+    echo "ceilometer event_pipeline.yaml file does not exist"
+    exit 1
+fi
+if [ -e /etc/nova/nova.conf ]; then
+    if ! grep -q '^notification_driver=messaging$' /etc/nova/nova.conf; then
+        sed -i -r 's/notification_driver=/notification_driver=messaging/g' /etc/nova/nova.conf
+        echo "modify nova config"
+        service nova-api restart
+    fi
+else
+    echo "nova.conf file does not exist"
+    exit 1
+fi
+exit 0
+END_TXT
+        chmod +x set_conf.sh
+        CONTROLLER_IP=$(sshpass -p r00tme ssh 2>/dev/null $ssh_opts root@${INSTALLER_IP} \
+             "fuel node | grep controller | cut -d '|' -f 5|xargs")
+        for node in $CONTROLLER_IP;do
+            scp $ssh_opts_cpu set_conf.sh "root@$node:"
+            ssh $ssh_opts_cpu "root@$node" './set_conf.sh > set_conf.log 2>&1 &'
+            sleep 1
+            scp $ssh_opts_cpu "root@$node:set_conf.log" set_conf_$node.log
+        done
+
+        if grep -q "modify the ceilometer config" set_conf_*.log ; then
+            NEED_TO_RESTORE_CEILOMETER=true
+        fi
+        if grep -q "modify nova config" set_conf_*.log ; then
+            NEED_TO_RESTORE_NOVA=true
         fi
 
-        echo "INSTALLER_TYPE set to 'local'. Assuming SSH keys already exchanged with $COMPUTE_HOST"
+        echo "waiting service restart..."
+        sleep 60
+    fi
+}
+
+restore_test_env() {
+    #TODO delete it when fuel support the configuration
+    if [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        if ! ($NEED_TO_RESTORE_CEILOMETER || $NEED_TO_RESTORE_NOVA) ; then
+            echo "Don't need to restore config"
+            exit 0
+        fi
+
+        echo "restore the configuration..."
+        cat > restore_conf.sh << 'END_TXT'
+#!/bin/bash
+if @NEED_TO_RESTORE_CEILOMETER@ ; then
+    if [ -e /etc/ceilometer/event_pipeline.yaml ]; then
+        if grep -q '^ *- notifier://?topic=alarm.all$' /etc/ceilometer/event_pipeline.yaml; then
+            sed -i 's|- notifier://?topic=alarm.all|- notifier://|' /etc/ceilometer/event_pipeline.yaml
+            service ceilometer-agent-notification restart
+        fi
+    else
+        echo "ceilometer event_pipeline.yaml file does not exist"
+        exit 1
     fi
+fi
+if @NEED_TO_RESTORE_NOVA@ ; then
+    if [ -e /etc/nova/nova.conf ]; then
+        if grep -q '^notification_driver=messaging$' /etc/nova/nova.conf; then
+            sed -i -r 's/notification_driver=messaging/notification_driver=/g' /etc/nova/nova.conf
+            service nova-api restart
+        fi
+    else
+        echo "nova.conf file does not exist"
+        exit 1
+    fi
+fi
+exit 0
+END_TXT
+        sed -i -e "s/@NEED_TO_RESTORE_CEILOMETER@/$NEED_TO_RESTORE_CEILOMETER/" restore_conf.sh
+        sed -i -e "s/@NEED_TO_RESTORE_NOVA@/$NEED_TO_RESTORE_NOVA/" restore_conf.sh
+        chmod +x restore_conf.sh
+        for node in $CONTROLLER_IP;do
+            scp $ssh_opts_cpu restore_conf.sh "root@$node:"
+            ssh $ssh_opts_cpu "root@$node" './restore_conf.sh > restore_conf.log 2>&1 &'
+        done
+
+        echo "waiting service restart..."
+        sleep 60
+    fi
+}
+
+get_compute_host_info() {
+    # get computer host info which VM boot in
+    COMPUTE_HOST=$(openstack $as_doctor_user server show $VM_NAME |
+                   grep "OS-EXT-SRV-ATTR:host" | awk '{ print $4 }')
+    compute_host_in_undercloud=${COMPUTE_HOST%%.*}
+    if [[ -z "$COMPUTE_HOST" ]] ; then
+        echo "ERROR: failed to get compute hostname"
+        exit 1
+    fi
+
+    if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
+        COMPUTE_USER=${COMPUTE_USER:-heat-admin}
+        COMPUTE_IP=$(sudo ssh $ssh_opts $INSTALLER_IP \
+             "source stackrc; \
+             nova show $compute_host_in_undercloud \
+             | awk '/ ctlplane network /{print \$5}'")
+    elif [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        COMPUTE_USER=${COMPUTE_USER:-root}
+        node_id=$(echo $compute_host_in_undercloud | cut -d "-" -f 2)
+        COMPUTE_IP=$(sshpass -p r00tme ssh 2>/dev/null $ssh_opts root@${INSTALLER_IP} \
+             "fuel node|awk -F '|' -v id=$node_id '{if (\$1 == id) print \$5}' |xargs")
+    elif [[ "$INSTALLER_TYPE" == "local" ]] ; then
+        COMPUTE_USER=${COMPUTE_USER:-$(whoami)}
+        COMPUTE_IP=$(getent hosts "$COMPUTE_HOST" | awk '{ print $1 }')
+    fi
+
+    if [[ -z "$COMPUTE_IP" ]]; then
+        echo "ERROR: Could not resolve $COMPUTE_HOST. Either manually set COMPUTE_IP or enable DNS resolution."
+        exit 1
+    fi
+    echo "COMPUTE_HOST=$COMPUTE_HOST"
+    echo "COMPUTE_IP=$COMPUTE_IP"
 
     # verify connectivity to target compute host
     ping -c 1 "$COMPUTE_IP"
+    if [[ $? -ne 0 ]] ; then
+        echo "ERROR: can not ping to computer host"
+        exit 1
+    fi
+
+    # verify ssh to target compute host
+    ssh $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP" 'exit'
+    if [[ $? -ne 0 ]] ; then
+        echo "ERROR: can not ssh to computer host"
+        exit 1
+    fi
+}
+
+get_consumer_ip() {
+    local get_consumer_command="ip route get $COMPUTE_IP | awk '/ src /{print \$NF}'"
+    if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
+        CONSUMER_IP=$(sudo ssh $ssh_opts root@$INSTALLER_IP \
+                      "$get_consumer_command")
+    elif [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+        CONSUMER_IP=$(sudo sshpass -p r00tme ssh $ssh_opts root@${INSTALLER_IP} \
+                      "$get_consumer_command")
+    elif [[ "$INSTALLER_TYPE" == "local" ]] ; then
+        CONSUMER_IP=`$get_consumer_command`
+    fi
+    echo "CONSUMER_IP=$CONSUMER_IP"
+
+    if [[ -z "$CONSUMER_IP" ]]; then
+        echo "ERROR: Could not get CONSUMER_IP."
+        exit 1
+    fi
 }
 
 download_image() {
@@ -74,26 +257,45 @@ download_image() {
 }
 
 register_image() {
-    glance image-list | grep -q " $IMAGE_NAME " && return 0
-    glance image-create --name "$IMAGE_NAME" \
-                        --visibility public \
-                        --disk-format "$IMAGE_FORMAT" \
-                        --container-format bare \
-                        --file "$IMAGE_FILE"
+    openstack image list | grep -q " $IMAGE_NAME " && return 0
+    openstack image create "$IMAGE_NAME" \
+                           --public \
+                           --disk-format "$IMAGE_FORMAT" \
+                           --container-format bare \
+                           --file "$IMAGE_FILE"
+}
+
+create_test_user() {
+    openstack project list | grep -q " $DOCTOR_PROJECT " || {
+        openstack project create "$DOCTOR_PROJECT"
+    }
+    openstack user list | grep -q " $DOCTOR_USER " || {
+        openstack user create "$DOCTOR_USER" --password "$DOCTOR_PW" \
+                              --project "$DOCTOR_PROJECT"
+    }
+    openstack user role list "$DOCTOR_USER" --project "$DOCTOR_PROJECT" \
+    | grep -q " $DOCTOR_ROLE " || {
+        openstack role add "$DOCTOR_ROLE" --user "$DOCTOR_USER" \
+                           --project "$DOCTOR_PROJECT"
+    }
 }
 
 boot_vm() {
-    nova list | grep -q " $VM_NAME " && return 0
-    nova boot --flavor "$VM_FLAVOR" \
-              --image "$IMAGE_NAME" \
-              "$VM_NAME"
+    # test VM done with test user, so can test non-admin
+    openstack $as_doctor_user server list | grep -q " $VM_NAME " && return 0
+    openstack $as_doctor_user server create --flavor "$VM_FLAVOR" \
+                            --image "$IMAGE_NAME" \
+                            "$VM_NAME"
     sleep 1
 }
 
 create_alarm() {
-    ceilometer alarm-list | grep -q " $ALARM_NAME " && return 0
-    vm_id=$(nova list | grep " $VM_NAME " | awk '{print $2}')
-    ceilometer alarm-event-create --name "$ALARM_NAME" \
+    # get vm_id as test user
+    ceilometer $as_doctor_user alarm-list | grep -q " $ALARM_NAME " && return 0
+    vm_id=$(openstack $as_doctor_user server list | grep " $VM_NAME " | awk '{print $2}')
+    # TODO(r-mibu): change notification endpoint from localhost to the consumer
+    # IP address (functest container).
+    ceilometer $as_doctor_user alarm-event-create --name "$ALARM_NAME" \
         --alarm-action "http://localhost:$CONSUMER_PORT/failure" \
         --description "VM failure" \
         --enabled True \
@@ -103,61 +305,172 @@ create_alarm() {
         -q "traits.state=string::error; traits.instance_id=string::$vm_id"
 }
 
+print_log() {
+    log_file=$1
+    echo "$log_file:"
+    sed -e 's/^/    /' "$log_file"
+}
+
 start_monitor() {
     pgrep -f "python monitor.py" && return 0
-    sudo python monitor.py "$COMPUTE_HOST" "$COMPUTE_IP" \
+    sudo -E python monitor.py "$COMPUTE_HOST" "$COMPUTE_IP" "$INSPECTOR_TYPE" \
         "http://127.0.0.1:$INSPECTOR_PORT/events" > monitor.log 2>&1 &
 }
 
 stop_monitor() {
     pgrep -f "python monitor.py" || return 0
     sudo kill $(pgrep -f "python monitor.py")
-    cat monitor.log
+    print_log monitor.log
+}
+
+congress_add_rule() {
+    name=$1
+    policy=$2
+    rule=$3
+
+    if ! openstack congress policy rule list $policy | grep -q -e "// Name: $name$" ; then
+        openstack congress policy rule create --name $name $policy "$rule"
+    fi
+}
+
+congress_del_rule() {
+    name=$1
+    policy=$2
+
+    if openstack congress policy rule list $policy | grep -q -e "^// Name: $name$" ; then
+        openstack congress policy rule delete $policy $name
+    fi
+}
+
+congress_setup_rules() {
+    congress_add_rule host_down classification \
+        'host_down(host) :-
+            doctor:events(hostname=host, type="compute.host.down", status="down")'
+
+    congress_add_rule active_instance_in_host classification \
+        'active_instance_in_host(vmid, host) :-
+            nova:servers(id=vmid, host_name=host, status="ACTIVE")'
+
+    congress_add_rule host_force_down classification \
+        'execute[nova:services.force_down(host, "nova-compute", "True")] :-
+            host_down(host)'
+
+    congress_add_rule error_vm_states classification \
+        'execute[nova:servers.reset_state(vmid, "error")] :-
+            host_down(host),
+            active_instance_in_host(vmid, host)'
 }
 
 start_inspector() {
-    pgrep -f "python inspector.py" && return 0
-    python inspector.py "$INSPECTOR_PORT" > inspector.log 2>&1 &
+    if [[ "$INSPECTOR_TYPE" == "sample" ]] ; then
+        pgrep -f "python inspector.py" && return 0
+        python inspector.py "$INSPECTOR_PORT" > inspector.log 2>&1 &
+    elif [[ "$INSPECTOR_TYPE" == "congress" ]] ; then
+        nova_api_min_version="2.11"
+        nova_api_version=$(openstack congress datasource list | \
+                           grep nova | grep -Po "(?<='api_version': ')[^']*")
+        [[ -z $nova_api_version ]] && nova_api_version="2.0"
+        if [[ "$nova_api_version" < "$nova_api_min_version" ]]; then
+            echo "ERROR: Congress Nova datasource API version < $nova_api_min_version ($nova_api_version)"
+            exit 1
+        fi
+        openstack congress driver list | grep -q " doctor "
+        openstack congress datasource list | grep -q " doctor " || {
+            openstack congress datasource create doctor doctor
+        }
+        congress_setup_rules
+    fi
 }
 
 stop_inspector() {
-    pgrep -f "python inspector.py" || return 0
-    kill $(pgrep -f "python inspector.py")
-    cat inspector.log
+    if [[ "$INSPECTOR_TYPE" == "sample" ]] ; then
+        pgrep -f "python inspector.py" || return 0
+        kill $(pgrep -f "python inspector.py")
+        print_log inspector.log
+    elif [[ "$INSPECTOR_TYPE" == "congress" ]] ; then
+        congress_del_rule host_force_down classification
+        congress_del_rule error_vm_states classification
+        congress_del_rule active_instance_in_host classification
+        congress_del_rule host_down classification
+    fi
 }
 
 start_consumer() {
     pgrep -f "python consumer.py" && return 0
     python consumer.py "$CONSUMER_PORT" > consumer.log 2>&1 &
+
+    # NOTE(r-mibu): create tunnel to the controller nodes, so that we can
+    # avoid some network problems dpends on infra and installers.
+    # This tunnel will be terminated by stop_consumer() or after 10 mins passed.
+    if [[ "$INSTALLER_TYPE" != "local" ]] ; then
+        if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
+            CONTROLLER_IPS=$(sudo ssh $ssh_opts $INSTALLER_IP \
+                             "source stackrc; \
+                             nova list | grep ' overcloud-controller-[0-9] ' \
+                             | sed -e 's/^.*ctlplane=//' -e 's/ *|\$//'")
+        elif [[ "$INSTALLER_TYPE" == "fuel" ]] ; then
+            CONTROLLER_IPS=$(sshpass -p r00tme ssh 2>/dev/null $ssh_opts root@${INSTALLER_IP} \
+                            "fuel node | grep controller | cut -d '|' -f 5|xargs")
+        fi
+
+        if [[ -z "$CONTROLLER_IPS" ]]; then
+            echo "ERROR: Could not get CONTROLLER_IPS."
+            exit 1
+        fi
+        for ip in $CONTROLLER_IPS
+        do
+            forward_rule="-R $CONSUMER_PORT:localhost:$CONSUMER_PORT"
+            tunnel_command="sudo ssh $ssh_opts_cpu $COMPUTE_USER@$ip $forward_rule sleep 600"
+            $tunnel_command > "ssh_tunnel.${ip}.log" 2>&1 < /dev/null &
+        done
+    fi
 }
 
 stop_consumer() {
     pgrep -f "python consumer.py" || return 0
     kill $(pgrep -f "python consumer.py")
-    cat consumer.log
+    print_log consumer.log
+
+    # NOTE(r-mibu): terminate tunnels to the controller nodes
+    if [[ "$INSTALLER_TYPE" != "local" ]] ; then
+        for ip in $CONTROLLER_IPS
+        do
+            forward_rule="-R $CONSUMER_PORT:localhost:$CONSUMER_PORT"
+            tunnel_command="sudo ssh $ssh_opts_cpu $COMPUTE_USER@$ip $forward_rule sleep 600"
+            kill $(pgrep -f "$tunnel_command")
+            print_log "ssh_tunnel.${ip}.log"
+        done
+    fi
 }
 
 wait_for_vm_launch() {
     echo "waiting for vm launch..."
-    while true
+
+    count=0
+    while [[ ${count} -lt 60 ]]
     do
-        state=$(nova list | grep " $VM_NAME " | awk '{print $6}')
+        state=$(openstack $as_doctor_user server list | grep " $VM_NAME " | awk '{print $6}')
         [[ "$state" == "ACTIVE" ]] && return 0
+        [[ "$state" == "ERROR" ]] && echo "vm state is ERROR" && exit 1
+        count=$(($count+1))
         sleep 1
     done
+    echo "ERROR: time out while waiting for vm launch"
+    exit 1
 }
 
 inject_failure() {
     echo "disabling network of compute host [$COMPUTE_HOST] for 3 mins..."
     cat > disable_network.sh << 'END_TXT'
 #!/bin/bash -x
-dev=$(sudo ip route | awk '/^default/{print $5}')
+dev=$(sudo ip a | awk '/ @COMPUTE_IP@\//{print $7}')
 sleep 1
 sudo ip link set $dev down
 sleep 180
 sudo ip link set $dev up
 sleep 1
 END_TXT
+    sed -i -e "s/@COMPUTE_IP@/$COMPUTE_IP/" disable_network.sh
     chmod +x disable_network.sh
     scp $ssh_opts_cpu disable_network.sh "$COMPUTE_USER@$COMPUTE_IP:"
     ssh $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP" 'nohup ./disable_network.sh > disable_network.log 2>&1 &'
@@ -166,10 +479,31 @@ END_TXT
 calculate_notification_time() {
     detected=$(grep "doctor monitor detected at" monitor.log | awk '{print $5}')
     notified=$(grep "doctor consumer notified at" consumer.log | awk '{print $5}')
+    if ! grep -q "doctor consumer notified at" consumer.log ; then
+        echo "ERROR: consumer hasn't received fault notification."
+        exit 1
+    fi
     echo "$notified $detected" | \
         awk '{d = $1 - $2; if (d < 1 && d > 0) print d " OK"; else print d " NG"}'
 }
 
+check_host_status() {
+    expected_state=$1
+
+    host_status_line=$(openstack $as_doctor_user --os-compute-api-version 2.16 \
+                       server show $VM_NAME | grep "host_status")
+    host_status=$(echo $host_status_line | awk '{print $4}')
+    if [ -z "$host_status" ] ; then
+        echo "ERROR: host_status not reported by: nova show $VM_NAME"
+        exit 1
+    elif [[ "$expected_state" =~ "$host_status" ]] ; then
+        echo "$VM_NAME showing host_status: $host_status"
+    else
+        echo "ERROR: host_status:$host_status not equal to expected_state: $expected_state"
+        exit 1
+    fi
+}
+
 cleanup() {
     set +e
     echo "cleanup..."
@@ -177,50 +511,72 @@ cleanup() {
     stop_inspector
     stop_consumer
 
+    echo "waiting disabled compute host back to be enabled..."
     python ./nova_force_down.py "$COMPUTE_HOST" --unset
+    sleep 240
+    check_host_status "UP"
+    scp $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP:disable_network.log" .
+    print_log disable_network.log
+
+    openstack $as_doctor_user server list | grep -q " $VM_NAME " && openstack $as_doctor_user server delete "$VM_NAME"
     sleep 1
-    nova delete "$VM_NAME"
-    sleep 1
-    alarm_id=$(ceilometer alarm-list | grep " $ALARM_NAME " | awk '{print $2}')
+    alarm_id=$(ceilometer $as_doctor_user alarm-list | grep " $ALARM_NAME " | awk '{print $2}')
     sleep 1
-    [ -n "$alarm_id" ] && ceilometer alarm-delete "$alarm_id"
+    [ -n "$alarm_id" ] && ceilometer $as_doctor_user alarm-delete "$alarm_id"
     sleep 1
-    image_id=$(glance image-list | grep " $IMAGE_NAME " | awk '{print $2}')
+
+    image_id=$(openstack image list | grep " $IMAGE_NAME " | awk '{print $2}')
     sleep 1
-    [ -n "$image_id" ] && glance image-delete "$image_id"
-    #TODO: add host status check via nova admin api
-    echo "waiting disabled compute host back to be enabled..."
-    sleep 180
-    ssh $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP" \
-        "[ -e disable_network.log ] && cat disable_network.log"
+    [ -n "$image_id" ] && openstack image delete "$image_id"
+    openstack role remove "$DOCTOR_ROLE" --user "$DOCTOR_USER" \
+                              --project "$DOCTOR_PROJECT"
+    openstack project delete "$DOCTOR_PROJECT"
+    openstack user delete "$DOCTOR_USER"
+
+    restore_test_env
 }
 
 
 echo "Note: doctor/tests/run.sh has been executed."
 
-prepare_compute_ssh
-
 trap cleanup EXIT
 
+echo "preparing test env..."
+get_installer_ip
+prepare_ssh_to_cloud
+prepare_test_env
+
 echo "preparing VM image..."
 download_image
 register_image
 
-echo "starting doctor sample components..."
-start_monitor
-start_inspector
-start_consumer
+echo "creating test user..."
+create_test_user
 
-echo "creating VM and alarm..."
+echo "creating VM..."
 boot_vm
-create_alarm
 wait_for_vm_launch
+openstack $as_doctor_user server show $VM_NAME
+
+echo "get computer host info..."
+get_compute_host_info
+
+echo "creating alarm..."
+#TODO: change back to use, network problems depends on infra and installers 
+#get_consumer_ip
+create_alarm
+
+echo "starting doctor sample components..."
+start_inspector
+start_monitor
+start_consumer
 
 sleep 60
 echo "injecting host failure..."
 inject_failure
-sleep 10
+sleep 60
 
+check_host_status "(DOWN|UNKNOWN)"
 calculate_notification_time
 
 echo "done"