summaryrefslogtreecommitdiffstats
path: root/tests/run.sh
blob: 7008650211eb3ab55ab3044549b3b7d0ea47acbd (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
#!/bin/bash -e
##############################################################################
# Copyright (c) 2016 NEC Corporation and others.
#
# All rights reserved. This program and the accompanying materials
# are made available under the terms of the Apache License, Version 2.0
# which accompanies this distribution, and is available at
# http://www.apache.org/licenses/LICENSE-2.0
##############################################################################

# Configuration

[[ "${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
#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
VM_FLAVOR=m1.tiny
ALARM_NAME=doctor_alarm1
INSPECTOR_PORT=12345
CONSUMER_PORT=12346
DOCTOR_USER=doctor
DOCTOR_PW=doctor
DOCTOR_PROJECT=doctor
#TODO: change back to `_member_` when JIRA DOCTOR-55 is done
DOCTOR_ROLE=admin
PROFILER_TYPE=${PROFILER_TYPE:-none}

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
                --os-tenant-name $DOCTOR_PROJECT"


# Functions

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%%.*}
    die_if_not_set $LINENO COMPUTE_HOST "Failed to get compute hostname"

    if is_installer apex; then
        COMPUTE_USER=${COMPUTE_USER:-heat-admin}
        COMPUTE_IP=$(sudo ssh $ssh_opts $INSTALLER_IP \
             "source stackrc; \
             nova show $compute_host_in_undercloud \
             | awk '/ ctlplane network /{print \$5}'")
    elif is_installer fuel; then
        COMPUTE_USER=${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=$node_id '{if (\$1 == id) print \$5}' |xargs")
    elif is_installer local; then
        COMPUTE_USER=${COMPUTE_USER:-$(whoami)}
        COMPUTE_IP=$(getent hosts "$COMPUTE_HOST" | awk '{ print $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
        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
        die $LINENO "Can not ssh to computer host"
    fi
}

get_consumer_ip() {
    local get_consumer_command="ip route get $COMPUTE_IP | awk '/ src /{print \$NF}'"
    if is_installer apex; then
        CONSUMER_IP=$(sudo ssh $ssh_opts root@$INSTALLER_IP \
                      "$get_consumer_command")
    elif is_installer fuel; then
        CONSUMER_IP=$(sudo sshpass -p r00tme ssh $ssh_opts root@${INSTALLER_IP} \
                      "$get_consumer_command")
    elif is_installer local; then
        CONSUMER_IP=`$get_consumer_command`
    fi
    echo "CONSUMER_IP=$CONSUMER_IP"

    die_if_not_set $LINENO CONSUMER_IP "Could not get CONSUMER_IP."
}

download_image() {
    #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() {
    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() {
    # 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() {
    # 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:$CONSUMER_PORT/failure" \
        --description "VM failure" \
        --enabled True \
        --repeat-actions False \
        --severity "moderate" \
        --event-type compute.instance.update \
        -q "traits.state=string::error; traits.instance_id=string::$vm_id"
}

start_monitor() {
    pgrep -f "python monitor.py" && return 0
    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")
}

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 ! is_installer local; then
        if is_installer 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 is_installer fuel; then
            get_controller_ips
        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"
            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")

    # NOTE(r-mibu): terminate tunnels to the controller nodes
    if ! is_installer 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")
        done
    fi
}

wait_for_vm_launch() {
    echo "waiting for vm launch..."

    count=0
    while [[ ${count} -lt 60 ]]
    do
        state=$(openstack $as_doctor_user server list | grep " $VM_NAME " | awk '{print $6}')
        if [[ "$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
            openstack $as_doctor_user server show $VM_NAME
            die $LINENO "vm state is ERROR"
        fi
        count=$(($count+1))
        sleep 1
    done
    die $LINENO "Time out while waiting for VM launch"
}

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 a | awk '/ @COMPUTE_IP@\//{print $7}')
[[ -n "$dev" ]] || dev=$(sudo ip a | awk '/ @COMPUTE_IP@\//{print $5}')
sleep 1
sudo ip link set $dev down
echo "doctor set link down at" $(date "+%s.%N")
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@$COMPUTE_IP" 'nohup ./disable_network.sh > disable_network.log 2>&1 &'
    # use host time to get rid of potential time sync deviation between nodes
    triggered=$(date "+%s.%N")
}

calculate_notification_time() {
    if ! grep -q "doctor consumer notified at" consumer.log ; then
        die $LINENO "Consumer hasn't received fault notification."
    fi

    #keep 'at' as the last keyword just before the value, and
    #use regex to get value instead of the fixed column
    detected=$(grep "doctor monitor detected at" monitor.log |\
               sed -e "s/^.* at //")
    notified=$(grep "doctor consumer notified at" consumer.log |\
               sed -e "s/^.* at //")

    echo "$notified $detected" | \
        awk '{
            d = $1 - $2;
            if (d < 1 && d > 0) { print d " OK"; exit 0 }
            else { print d " NG"; exit 1 }
        }'
}

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}')
    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
        die $LINENO "host_status:$host_status not equal to expected_state: $expected_state"
    fi
}

unset_forced_down_hosts() {
    for host in $(openstack compute service list --service nova-compute \
                  -f value -c Host -c State | sed -n -e '/down$/s/ *down$//p')
    do
        # TODO (r-mibu): make sample inspector use keystone v3 api
        OS_AUTH_URL=${OS_AUTH_URL/v3/v2.0} \
        python ./nova_force_down.py $host --unset
    done

    echo "waiting disabled compute host back to be enabled..."
    wait_until 'openstack compute service list --service nova-compute
                -f value -c State | grep -q down' 240 5
}

collect_logs() {
    unset_forced_down_hosts
    # TODO: We need to make sure the target compute host is back to IP
    #       reachable. wait_ping() will be added by tojuvone .
    sleep 110
    scp $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP:disable_network.log" .

    # TODO(yujunz) collect other logs, e.g. nova, aodh
}

run_profiler() {
    if [[ "$PROFILER_TYPE" == "poc" ]]; then
        linkdown=$(grep "doctor set link down at " disable_network.log |\
                  sed -e "s/^.* at //")
        vmdown=$(grep "doctor mark vm.* error at" inspector.log |tail -n 1 |\
                 sed -e "s/^.* at //")
        hostdown=$(grep "doctor mark host.* down at" inspector.log |\
                 sed -e "s/^.* at //")

        # TODO(yujunz) check the actual delay to verify time sync status
        # expected ~1s delay from $trigger to $linkdown
        relative_start=${linkdown}
        export DOCTOR_PROFILER_T00=$(python -c \
          "print(int(($linkdown-$relative_start)*1000))")
        export DOCTOR_PROFILER_T01=$(python -c \
          "print(int(($detected-$relative_start)*1000))")
        export DOCTOR_PROFILER_T03=$(python -c \
          "print(int(($vmdown-$relative_start)*1000))")
        export DOCTOR_PROFILER_T04=$(python -c \
          "print(int(($hostdown-$relative_start)*1000))")
        export DOCTOR_PROFILER_T09=$(python -c \
          "print(int(($notified-$relative_start)*1000))")

        python profiler-poc.py >doctor_profiler.log 2>&1
    fi
}

cleanup() {
    set +e
    echo "cleanup..."
    stop_monitor
    stop_inspector
    stop_consumer

    unset_forced_down_hosts
    # TODO: We need to make sure the target compute host is back to IP
    #       reachable. wait_ping() will be added by tojuvone .
    sleep 110
    scp $ssh_opts_cpu "$COMPUTE_USER@$COMPUTE_IP:disable_network.log" .

    openstack $as_doctor_user server list | grep -q " $VM_NAME " && openstack $as_doctor_user server delete "$VM_NAME"
    sleep 1
    alarm_id=$(ceilometer $as_doctor_user alarm-list | grep " $ALARM_NAME " | awk '{print $2}')
    sleep 1
    [ -n "$alarm_id" ] && ceilometer $as_doctor_user alarm-delete "$alarm_id"
    sleep 1

    image_id=$(openstack image list | grep " $IMAGE_NAME " | awk '{print $2}')
    sleep 1
    #if an existing image was used, there's no need to remove it here
    if [[ "$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"
    openstack user delete "$DOCTOR_USER"

    cleanup_installer
    cleanup_inspector
}

# Main process

echo "Note: doctor/tests/run.sh has been executed."
git log --oneline -1 || true   # ignore even you don't have git installed

trap cleanup EXIT

source $TOP_DIR/functions-common
source $TOP_DIR/lib/installer
source $TOP_DIR/lib/inspector

setup_installer

echo "preparing VM image..."
download_image
register_image

echo "creating test user..."
create_test_user

echo "creating VM..."
boot_vm
wait_for_vm_launch

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 60

check_host_status "(DOWN|UNKNOWN)"
calculate_notification_time
collect_logs
run_profiler

echo "done"