Merge pull request #2727 from weaveworks/issues/2397-v2-plugin

Create a V2 Docker plugin

test/090_docker_restart_policy_2_test.sh

@@ -6,22 +6,20 @@ start_suite "Test docker restart policy"
 
 weave_on $HOST1 launch
 
-assert "docker_on $HOST1 inspect -f '{{.HostConfig.RestartPolicy.Name}}' weave weaveproxy weaveplugin" "always\nalways\nalways"
+assert "docker_on $HOST1 inspect -f '{{.HostConfig.RestartPolicy.Name}}' weave weaveproxy" "always\nalways"
 
 assert_raises "check_restart $HOST1 weave"
 assert_raises "check_restart $HOST1 weaveproxy"
-assert_raises "check_restart $HOST1 weaveplugin"
 
 # stop + launch tests that restart policy changes result
 # in the old containers being removed and new ones created
 weave_on $HOST1 stop
 weave_on $HOST1 launch --no-restart
 
-assert "docker_on $HOST1 inspect -f '{{.HostConfig.RestartPolicy.Name}}' weave weaveproxy weaveplugin" "no\nno\nno"
+assert "docker_on $HOST1 inspect -f '{{.HostConfig.RestartPolicy.Name}}' weave weaveproxy" "no\nno"
 
 assert_raises "! check_restart $HOST1 weave"
 assert_raises "! check_restart $HOST1 weaveproxy"
-assert_raises "! check_restart $HOST1 weaveplugin"
 
 # Relaunch to prevent the `weave stop` in `end_suite`
 # timing out trying to remove the plugin network

test/195_no_multicast_route_1_test.sh

@@ -12,9 +12,8 @@ start_suite "--no-multicast-route operation"
 $SSH $HOST1 docker network rm testmcasttrue testmcastfalse >/dev/null 2>&1 || true
 
 # Ensure containers run either way have no multicast route
-weave_on $HOST1 launch-router
+weave_on $HOST1 launch-router --plugin
 weave_on $HOST1 launch-proxy --no-multicast-route
-weave_on $HOST1 launch-plugin # plugin defaults to multicast off except for 'net=weave'
 
 docker_on $HOST1 run --name c1 -dt $SMALL_IMAGE /bin/sh
 weave_on $HOST1 attach --no-multicast-route c1

test/335_rmpeer_3_test.sh

@@ -14,7 +14,10 @@ weave_on $HOST3 launch $HOST1
 weave_on $HOST3 prime
 
 # nuke 1st peer
-docker_on $HOST1 kill weave
+# NOTE: docker-kill hangs (https://github.com/docker/docker/issues/31447), so we
+# kill directly the weaver process instead.
+WEAVER_PID=$(container_pid $HOST1 weave)
+run_on $HOST1 "sudo kill -9 $WEAVER_PID"
 
 # transfer its space to the 2nd
 weave_on $HOST2 rmpeer $PEER1

test/600_proxy_docker_py.sh

@@ -3,7 +3,7 @@
 
 . "$(dirname "$0")/config.sh"
 
-IMAGE="joffrey/docker-py:1.8.1"
+IMAGE="joffrey/docker-py:1.9.0-rc2"
 
 docker_py_test() {
     SHARD=$1

test/620_proxy_entrypoint_command_test.sh

@@ -15,7 +15,7 @@ run_container() {
 }
 
 start_suite "Proxy uses correct entrypoint and command with weavewait"
-weave_on $HOST1 launch-proxy
+weave_on $HOST1 launch-proxy --no-restart
 
 build_image check-ethwe-up '["grep"]' '["^1$", "/sys/class/net/ethwe/carrier"]'
 run_container "check-ethwe-up"
@@ -27,7 +27,11 @@ build_image false '["/bin/false"]' ''
 run_container "--entrypoint='grep' false ^1$ /sys/class/net/ethwe/carrier"
 
 weave_on $HOST1 launch-router --ipalloc-range 10.2.2.0/24
-docker_on $HOST1 kill weaveproxy
+# NOTE: docker-kill hangs (https://github.com/docker/docker/issues/31447), so we
+# kill directly the weaveproxy process instead.
+WEAVEPROXY_PID=$(container_pid $HOST1 weaveproxy)
+$SSH $HOST1 "ps aux | grep weaveproxy"
+run_on $HOST1 "sudo kill -9 $WEAVEPROXY_PID"
 weave_on $HOST1 launch-proxy
 
 assert_raises "proxy docker_on $HOST1 run check-ethwe-up"

test/850_plugin_v2_cross_hosts_2_test.sh

@@ -0,0 +1,128 @@
+#! /bin/bash
+
+. "$(dirname "$0")/config.sh"
+
+PLUGIN_NAME="weave-ci-registry:5000/weaveworks/net-plugin"
+HOST1_IP=$($SSH $HOST1 getent ahosts $HOST1 | grep "RAW" | cut -d' ' -f1)
+SERVICE="weave-850-service"
+NETWORK="weave-850-network"
+
+setup_master() {
+    # Setup Docker image registry on $HOST1
+    docker_on $HOST1 run -p 5000:5000 -d registry:2
+
+    # Build plugin-v2 on $HOST1, because Circle "CI" runs only ancient
+    # version of Docker which does not support v2 plugins.
+    build_plugin_img="buildpluginv2-ci"
+    rsync -az -e "$SSH" "$(dirname $0)/../prog/net-plugin" $HOST1:~/
+    $SSH $HOST1<<EOF
+        docker rmi $build_plugin_img 2>/dev/null
+        docker plugin disable $PLUGIN_NAME >/dev/null
+        docker plugin remove $PLUGIN_NAME 2>/dev/null
+
+        WORK_DIR=$(mktemp -d)
+        mkdir -p \${WORK_DIR}/rootfs
+
+        docker create --name=$build_plugin_img weaveworks/weave true
+        docker export $build_plugin_img | tar -x -C \${WORK_DIR}/rootfs
+        cp \${HOME}/net-plugin/launch.sh \${WORK_DIR}/rootfs/home/weave/launch.sh
+        cp \${HOME}/net-plugin/config.json \${WORK_DIR}
+        docker plugin create $PLUGIN_NAME \${WORK_DIR}
+
+        echo "$HOST1_IP weave-ci-registry" | sudo tee -a /etc/hosts
+        docker plugin push $PLUGIN_NAME
+
+        # Start Swarm Manager and enable the plugin
+        docker swarm init --advertise-addr=$HOST1_IP
+
+        [ -n "$COVERAGE" ] && docker plugin set $PLUGIN_NAME EXTRA_ARGS="-test.coverprofile=/home/weave/cover.prof --"
+        #docker plugin set $PLUGIN_NAME WEAVE_PASSWORD="foobar"
+        docker plugin enable $PLUGIN_NAME
+EOF
+}
+
+setup_worker() {
+    $SSH $HOST2<<EOF
+        echo "$HOST1_IP weave-ci-registry" | sudo tee -a /etc/hosts
+        ping -nq -W 2 -c 1 weave-ci-registry
+        docker swarm join --token "$1" "${HOST1_IP}:2377"
+        docker plugin install --disable --grant-all-permissions $PLUGIN_NAME
+
+        [ -n "$COVERAGE" ] && docker plugin set $PLUGIN_NAME EXTRA_ARGS="-test.coverprofile=/home/weave/cover.prof --"
+        #docker plugin set $PLUGIN_NAME WEAVE_PASSWORD="foobar"
+        docker plugin enable $PLUGIN_NAME
+EOF
+}
+
+cleanup() {
+    for h in $@; do
+        $SSH $h<<EOF
+            sudo sed -i '/weave-ci-registry/d' /etc/hosts
+            docker service rm $SERVICE || true
+            docker swarm leave --force || true
+            docker network rm $NETWORK || true
+            docker plugin disable -f $PLUGIN_NAME || true
+            docker plugin remove -f $PLUGIN_NAME || true
+EOF
+    done
+}
+
+wait_for_network() {
+    for i in $(seq 10); do
+        if docker_on $1 network inspect $2 >/dev/null 2>&1; then
+            return
+        fi
+        echo "Waiting for \"$2\" network"
+        sleep 2
+    done
+    echo "Failed to wait for \"$2\" network" >&2
+    return 1
+}
+
+wait_for_service() {
+    for i in $(seq 60); do
+        replicas=$($SSH $1 docker service ls -f="name=$2" | awk '{print $4}' | grep -v REPLICAS)
+        if [ "$replicas" == "$3/$3" ]; then
+            return
+        fi
+        echo "Waiting for \"$2\" service"
+        sleep 2
+    done
+    echo "Failed to wait for \"$2\" service" >&2
+    return 1
+}
+
+start_suite "Test Docker plugin-v2"
+
+#cleanup $HOST1 $HOST2
+#exit 1
+
+setup_master
+setup_worker $($SSH $HOST1 docker swarm join-token --quiet worker)
+
+echo "Creating network and service..."
+
+# Create network and service
+$SSH $HOST1<<EOF
+    docker plugin ls
+    docker network create --driver="${PLUGIN_NAME}:latest" $NETWORK || true
+    # Otherwise no containers will be scheduled on host2
+    sleep 20
+    docker service create --name=$SERVICE --network=$NETWORK --replicas=2 nginx || true
+    journalctl -r -u docker.service -n30
+EOF
+
+wait_for_service $HOST1 $SERVICE 2
+
+C1=$($SSH $HOST1 weave ps | grep -v weave:expose | awk '{print $1}')
+C2_IP=$($SSH $HOST2 weave ps | grep -v weave:expose | awk '{print $3}' | cut -d/ -f1)
+
+assert_raises "exec_on $HOST1 $C1 $PING $C2_IP"
+
+# We do not test "weave {status,launch}", because the weave script does not detect
+# plugin-v2 if its name is prefixed with a registry name.
+
+# Failing to cleanup will make the rest of the tests to fail
+cleanup $HOST1 $HOST2
+
+end_suite