fix(lag): target only replica if rlog core+replicant

there seems to be race conditions related to some tests with sessions
hitting the core and the replicant alternately and rlog.

for intance, if there is some delay in this replication, a new
connection made to the replica with a just-created session in the core
may not have been replicated to the replicant, resulting in a test
failure if it expects the session to be present.

since such replication lags are inherent to the core-replicant
topology, we can try to target only the replicant to avoid seeing this
inconsistent view of the system during the tests.

.ci/docker-compose-file/docker-compose-emqx-cluster-rlog.override.yaml

@@ -11,17 +11,23 @@ x-default-emqx: &default-emqx
 services:
   emqx1:
     <<: *default-emqx
+    container_name: node1.emqx.io
     environment:
       - "EMQX_HOST=node1.emqx.io"
       - "EMQX_CLUSTER__DB_BACKEND=rlog"
       - "EMQX_CLUSTER__RLOG__ROLE=core"
       - "EMQX_CLUSTER__STATIC__SEEDS=[emqx@node1.emqx.io]"
+      - "EMQX_LISTENERS__TCP__DEFAULT__PROXY_PROTOCOL=false"
+      - "EMQX_LISTENERS__WS__DEFAULT__PROXY_PROTOCOL=false"
 
   emqx2:
     <<: *default-emqx
+    container_name: node2.emqx.io
     environment:
       - "EMQX_HOST=node2.emqx.io"
       - "EMQX_CLUSTER__DB_BACKEND=rlog"
       - "EMQX_CLUSTER__RLOG__ROLE=replicant"
       - "EMQX_CLUSTER__RLOG__CORE_NODES=emqx@node1.emqx.io"
       - "EMQX_CLUSTER__STATIC__SEEDS=[emqx@node1.emqx.io]"
+      - "EMQX_LISTENERS__TCP__DEFAULT__PROXY_PROTOCOL=false"
+      - "EMQX_LISTENERS__WS__DEFAULT__PROXY_PROTOCOL=false"

.ci/docker-compose-file/python/pytest.sh

@@ -6,16 +6,24 @@
 set -x
 set +e
 
-LB="haproxy"
+EMQX_TEST_DB_BACKEND=$1
+if [ "$EMQX_TEST_DB_BACKEND" = "rlog" ]
+then
+  # target only replica to avoid replication races
+  TARGET_HOST="node2.emqx.io"
+else
+  # use loadbalancer
+  TARGET_HOST="haproxy"
+fi
 
 apk update && apk add git curl
 git clone -b develop-4.0 https://github.com/emqx/paho.mqtt.testing.git /paho.mqtt.testing
 pip install pytest
 
-pytest -v /paho.mqtt.testing/interoperability/test_client/V5/test_connect.py -k test_basic --host "$LB"
+pytest -v /paho.mqtt.testing/interoperability/test_client/V5/test_connect.py -k test_basic --host "$TARGET_HOST"
 RESULT=$?
 
-pytest -v /paho.mqtt.testing/interoperability/test_client --host "$LB"
+pytest -v /paho.mqtt.testing/interoperability/test_client --host "$TARGET_HOST"
 RESULT=$(( RESULT + $? ))
 
 # pytest -v /paho.mqtt.testing/interoperability/test_cluster --host1 "node1.emqx.io" --host2 "node2.emqx.io"

.ci/docker-compose-file/scripts/run-emqx.sh

@@ -14,11 +14,27 @@ fi
   echo "EMQX_ZONES__DEFAULT__MQTT__MAX_TOPIC_ALIAS=10"
 } >> .ci/docker-compose-file/conf.cluster.env
 
-is_cluster_up() {
-  docker exec -i node1.emqx.io \
+is_node_up() {
+  local node
+  node="$1"
+  docker exec -i "$node" \
          bash -c "emqx eval \"['emqx@node1.emqx.io','emqx@node2.emqx.io'] = maps:get(running_nodes, ekka_cluster:info()).\"" > /dev/null 2>&1
 }
 
+is_node_listening() {
+  local node
+  node="$1"
+  docker exec -i "$node" \
+         emqx eval "ok = case gen_tcp:connect(\"localhost\", 1883, []) of {ok, P} -> gen_tcp:close(P), ok; _ -> exit(1) end." > /dev/null 2>&1
+}
+
+is_cluster_up() {
+  is_node_up node1.emqx.io && \
+    is_node_up node2.emqx.io && \
+    is_node_listening node1.emqx.io && \
+    is_node_listening node2.emqx.io
+}
+
 docker-compose \
   -f .ci/docker-compose-file/docker-compose-emqx-cluster.yaml \
   $CLUSTER_OVERRIDES \

.github/workflows/run_fvt_tests.yaml

@@ -99,10 +99,13 @@ jobs:
         ./.ci/docker-compose-file/scripts/run-emqx.sh
     - name: make paho tests
       run: |
-        if ! docker exec -i python /scripts/pytest.sh; then
+        if ! docker exec -i python /scripts/pytest.sh "${{ matrix.cluster_db_backend }}"; then
           echo "DUMP_CONTAINER_LOGS_BGN"
+          echo "============== haproxy =============="
           docker logs haproxy
+          echo "==============  node1  =============="
           docker logs node1.emqx.io
+          echo "==============  node2  =============="
           docker logs node2.emqx.io
           echo "DUMP_CONTAINER_LOGS_END"
           exit 1