Merge "remove print_log"
[doctor.git] / tests / run.sh
index 292a7ce..b6bfc0b 100755 (executable)
@@ -11,7 +11,8 @@
 [[ "${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
+#if an existing image name is provided in the enviroment, use that one
+IMAGE_NAME=${IMAGE_NAME:-cirros}
 IMAGE_FILE="${IMAGE_NAME}.img"
 IMAGE_FORMAT=qcow2
 VM_NAME=doctor_vm1
@@ -31,6 +32,7 @@ INSTALLER_IP=${INSTALLER_IP:-none}
 
 SUPPORTED_INSPECTOR_TYPES="sample congress"
 INSPECTOR_TYPE=${INSPECTOR_TYPE:-sample}
+TOP_DIR=$(cd $(dirname "$0") && pwd)
 
 ssh_opts="-o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no"
 as_doctor_user="--os-username $DOCTOR_USER --os-password $DOCTOR_PW
@@ -60,10 +62,7 @@ get_installer_ip() {
     fi
 
     if [[ "$INSTALLER_TYPE" != "local" ]] ; then
-        if [[ -z "$INSTALLER_IP" ]] ; then
-            echo "ERROR: no installer ip"
-            exit 1
-        fi
+        die_if_not_set $LINENO INSTALLER_IP "No installer IP"
     fi
 }
 
@@ -189,10 +188,7 @@ get_compute_host_info() {
     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
+    die_if_not_set $LINENO COMPUTE_HOST "Failed to get compute hostname"
 
     if [[ "$INSTALLER_TYPE" == "apex" ]] ; then
         COMPUTE_USER=${COMPUTE_USER:-heat-admin}
@@ -210,25 +206,20 @@ get_compute_host_info() {
         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
+    die_if_not_set $LINENO COMPUTE_IP "Could not resolve $COMPUTE_HOST. Either manually set COMPUTE_IP or enable DNS resolution."
     echo "COMPUTE_HOST=$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
+        die $LINENO "Can not ping to computer host"
     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
+        die $LINENO "Can not ssh to computer host"
     fi
 }
 
@@ -245,15 +236,18 @@ get_consumer_ip() {
     fi
     echo "CONSUMER_IP=$CONSUMER_IP"
 
-    if [[ -z "$CONSUMER_IP" ]]; then
-        echo "ERROR: Could not get CONSUMER_IP."
-        exit 1
-    fi
+    die_if_not_set $LINENO CONSUMER_IP "Could not get CONSUMER_IP."
 }
 
 download_image() {
-    [ -e "$IMAGE_FILE" ] && return 0
-    wget "$IMAGE_URL" -o "$IMAGE_FILE"
+    #if a different name was provided for the image in the enviroment there's no need to download the image
+    use_existing_image=false
+    openstack image list | grep -q " $IMAGE_NAME " && use_existing_image=true
+
+    if [[ "$use_existing_image" == false ]] ; then
+        [ -e "$IMAGE_FILE" ] && return 0
+        wget "$IMAGE_URL" -o "$IMAGE_FILE"
+    fi
 }
 
 register_image() {
@@ -305,12 +299,6 @@ create_alarm() {
         -q "traits.state=string::error; traits.instance_id=string::$vm_id"
 }
 
-print_log() {
-    log_file=$1
-    echo "$log_file:"
-    sed -e 's/^/    /' "$log_file"
-}
-
 start_monitor() {
     pgrep -f "python monitor.py" && return 0
     sudo -E python monitor.py "$COMPUTE_HOST" "$COMPUTE_IP" "$INSPECTOR_TYPE" \
@@ -320,7 +308,6 @@ start_monitor() {
 stop_monitor() {
     pgrep -f "python monitor.py" || return 0
     sudo kill $(pgrep -f "python monitor.py")
-    print_log monitor.log
 }
 
 congress_add_rule() {
@@ -369,10 +356,11 @@ start_inspector() {
         nova_api_min_version="2.11"
         nova_api_version=$(openstack congress datasource list | \
                            grep nova | grep -Po "(?<='api_version': ')[^']*")
-        [[ -z $nova_api_version ]] && nova_api_version="2.0"
+        if ! is_set nova_api_version; then
+            nova_api_version="2.0"
+        fi
         if [[ "$nova_api_version" < "$nova_api_min_version" ]]; then
-            echo "ERROR: Congress Nova datasource API version < $nova_api_min_version ($nova_api_version)"
-            exit 1
+            die $LINENO "Congress Nova datasource API version < $nova_api_min_version ($nova_api_version)"
         fi
         openstack congress driver list | grep -q " doctor "
         openstack congress datasource list | grep -q " doctor " || {
@@ -386,7 +374,6 @@ stop_inspector() {
     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
@@ -413,10 +400,7 @@ start_consumer() {
                             "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
+        die_if_not_set $LINENO CONTROLLER_IPS "Could not get CONTROLLER_IPS."
         for ip in $CONTROLLER_IPS
         do
             forward_rule="-R $CONSUMER_PORT:localhost:$CONSUMER_PORT"
@@ -429,7 +413,6 @@ start_consumer() {
 stop_consumer() {
     pgrep -f "python consumer.py" || return 0
     kill $(pgrep -f "python consumer.py")
-    print_log consumer.log
 
     # NOTE(r-mibu): terminate tunnels to the controller nodes
     if [[ "$INSTALLER_TYPE" != "local" ]] ; then
@@ -438,7 +421,6 @@ stop_consumer() {
             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
 }
@@ -450,13 +432,19 @@ wait_for_vm_launch() {
     while [[ ${count} -lt 60 ]]
     do
         state=$(openstack $as_doctor_user server list | grep " $VM_NAME " | awk '{print $6}')
-        [[ "$state" == "ACTIVE" ]] && return 0
-        [[ "$state" == "ERROR" ]] && echo "vm state is ERROR" && exit 1
+        if [[ "$state" == "ACTIVE" ]]; then
+            # NOTE(cgoncalves): sleeping for a bit to stabilize
+            # See python-openstackclient/functional/tests/compute/v2/test_server.py:wait_for_status
+            sleep 5
+            return 0
+        fi
+        if [[ "$state" == "ERROR" ]]; then
+            die $LINENO "vm state is ERROR"
+        fi
         count=$(($count+1))
         sleep 1
     done
-    echo "ERROR: time out while waiting for vm launch"
-    exit 1
+    die $LINENO "Time out while waiting for VM launch"
 }
 
 inject_failure() {
@@ -477,14 +465,17 @@ 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 $5}')
+    detected=$(grep "doctor monitor detected at" monitor.log | awk '{print $10}')
+    notified=$(grep "doctor consumer notified at" consumer.log | awk '{print $10}')
     if ! grep -q "doctor consumer notified at" consumer.log ; then
-        echo "ERROR: consumer hasn't received fault notification."
-        exit 1
+        die $LINENO "Consumer hasn't received fault notification."
     fi
     echo "$notified $detected" | \
-        awk '{d = $1 - $2; if (d < 1 && d > 0) print d " OK"; else print d " NG"}'
+        awk '{
+            d = $1 - $2;
+            if (d < 1 && d > 0) { print d " OK"; exit 0 }
+            else { print d " NG"; exit 1 }
+        }'
 }
 
 check_host_status() {
@@ -493,14 +484,11 @@ check_host_status() {
     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
+    die_if_not_set $LINENO host_status "host_status not reported by: nova show $VM_NAME"
+    if [[ "$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
+        die $LINENO "host_status:$host_status not equal to expected_state: $expected_state"
     fi
 }
 
@@ -516,7 +504,6 @@ cleanup() {
     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
@@ -527,7 +514,10 @@ cleanup() {
 
     image_id=$(openstack image list | grep " $IMAGE_NAME " | awk '{print $2}')
     sleep 1
-    [ -n "$image_id" ] && openstack image delete "$image_id"
+    #if an existing image was used, there's no need to remove it here
+    if [[ "$use_existing_image" == false ]] ; then
+        [ -n "$image_id" ] && openstack image delete "$image_id"
+    fi
     openstack role remove "$DOCTOR_ROLE" --user "$DOCTOR_USER" \
                               --project "$DOCTOR_PROJECT"
     openstack project delete "$DOCTOR_PROJECT"
@@ -541,6 +531,8 @@ echo "Note: doctor/tests/run.sh has been executed."
 
 trap cleanup EXIT
 
+source $TOP_DIR/functions-common
+
 echo "preparing test env..."
 get_installer_ip
 prepare_ssh_to_cloud
@@ -556,13 +548,12 @@ create_test_user
 echo "creating VM..."
 boot_vm
 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 
+#TODO: change back to use, network problems depends on infra and installers
 #get_consumer_ip
 create_alarm