[dpu]: Add DPU database service (#17161)

Sub PRs:

sonic-net/sonic-host-services#84
#17191

Why I did it
According to the design, the database instances of DPU will be kept in the NPU host.

Microsoft ADO (number only): 25072889

How I did it
To follow the multiple ASIC design, I assume a new platform environment variable NUM_DPU will be defined in the /usr/share/sonic/device/$PLATFORM/platform_env.conf. Based on this number, NPU host will launch a corresponding number of instances for the DPU database.

Signed-off-by: Ze Gan <ganze718@gmail.com>

dockers/docker-database/Dockerfile.j2

@@ -32,7 +32,8 @@ RUN apt-get clean -y                                  && \
              s/^# unixsocket/unixsocket/;                \
              s/redis-server.sock/redis.sock/g;           \
              s/^client-output-buffer-limit pubsub [0-9]+mb [0-9]+mb [0-9]+/client-output-buffer-limit pubsub 0 0 0/; \
-             s/^notify-keyspace-events ""$/notify-keyspace-events AKE/ \
+             s/^notify-keyspace-events ""$/notify-keyspace-events AKE/; \
+             s/^databases [0-9]+$/databases 100/ \
             ' /etc/redis/redis.conf
 
 COPY ["supervisord.conf.j2", "/usr/share/sonic/templates/"]

dockers/docker-database/database_config.json.j2

@@ -2,9 +2,12 @@
     "INSTANCES": {
         "redis":{
             "hostname" : "{{HOST_IP}}",
-            "port" : 6379,
+            "port" : {{REDIS_PORT}},
-            "unix_socket_path" : "/var/run/redis{{NAMESPACE_ID}}/redis.sock",
+            "unix_socket_path" : "/var/run/redis{{DEV}}/redis.sock",
             "persistence_for_warm_boot" : "yes"
+            {% if DATABASE_TYPE is defined and DATABASE_TYPE != "" %}
+            ,"database_type": "{{DATABASE_TYPE}}"
+            {% endif %}
         },
         "redis_chassis":{
             "hostname" : "redis_chassis.server",
@@ -94,6 +97,30 @@
             "separator": ":",
             "instance" : "redis"
         }
+{% if DATABASE_TYPE is defined and DATABASE_TYPE == "dpudb" %}
+        ,
+        "DPU_APPL_DB" : {
+            "id" : 15,
+            "separator": ":",
+            "instance" : "redis",
+            "format": "proto"
+        },
+        "DPU_APPL_STATE_DB" : {
+            "id" : 16,
+            "separator": "|",
+            "instance" : "redis"
+        },
+        "DPU_STATE_DB" : {
+            "id" : 17,
+            "separator": "|",
+            "instance" : "redis"
+        },
+        "DPU_COUNTERS_DB" : {
+            "id" : 18,
+            "separator": ":",
+            "instance" : "redis"
+        }
+{% endif %}
     },
     "VERSION" : "1.0"
 }

dockers/docker-database/database_global.json.j2

@@ -1,9 +1,15 @@
 {% set namespace_cnt = NAMESPACE_COUNT|int %}
+{% if NUM_DPU is defined %}
+{% set dpu_cnt = NUM_DPU | int %}
+{% else %}
+{% set dpu_cnt = 0 %}
+{% endif %}
 {
     "INCLUDES" : [
         {
             "include" : "../../redis/sonic-db/database_config.json"
         },
+
 {% if namespace_cnt > 1 %}
 {% for ns in range(namespace_cnt) %}
         {
@@ -15,7 +21,22 @@
         },
 {% endif %}
 {% endfor %}
+{% endif %}
+
+{% if dpu_cnt > 0 %}
+{% for dpu in range(dpu_cnt) %}
+        {
+            "database_type" : "dpudb",
+            "include" : "../../redisdpu{{dpu}}/sonic-db/database_config.json"
+{% if dpu == dpu_cnt-1  %}
+        }
+{% else %}
+        },
+{% endif %}
+{% endfor %}
+{% endif %}
+
     ],
     "VERSION" : "1.0"
 }
-{% endif %}
+

dockers/docker-database/docker-database-init.sh

@@ -17,6 +17,18 @@ then
     host_ip=127.0.0.1
 fi
 
+redis_port=6379
+
+if [[ $DATABASE_TYPE == "dpudb" ]]; then
+    host_ip="169.254.200.254"
+    if ! ip -4 -o addr | awk '{print $4}' | grep $host_ip; then
+        host_ip=127.0.0.1
+    fi
+    DPU_ID=`echo $DEV | tr -dc '0-9'`
+    redis_port=`expr 6381 + $DPU_ID`
+fi
+
+
 REDIS_DIR=/var/run/redis$NAMESPACE_ID
 mkdir -p $REDIS_DIR/sonic-db
 mkdir -p /etc/supervisor/conf.d/
@@ -24,7 +36,7 @@ mkdir -p /etc/supervisor/conf.d/
 if [ -f /etc/sonic/database_config$NAMESPACE_ID.json ]; then
     cp /etc/sonic/database_config$NAMESPACE_ID.json $REDIS_DIR/sonic-db/database_config.json
 else
-    HOST_IP=$host_ip j2 /usr/share/sonic/templates/database_config.json.j2 > $REDIS_DIR/sonic-db/database_config.json
+    HOST_IP=$host_ip REDIS_PORT=$redis_port DATABASE_TYPE=$DATABASE_TYPE j2 /usr/share/sonic/templates/database_config.json.j2 > $REDIS_DIR/sonic-db/database_config.json
 fi
 
 # on VoQ system, we only publish redis_chassis instance and CHASSIS_APP_DB when
@@ -59,7 +71,7 @@ if [[ $DATABASE_TYPE == "chassisdb" ]]; then
 fi
 
 # copy/generate the database_global.json file if this is global database service in multi asic platform.
-if [[ $NAMESPACE_ID == "" ]] && [[ $NAMESPACE_COUNT -gt 1 ]]
+if [[ $DATABASE_TYPE == "" ]] && [[ $NAMESPACE_COUNT -gt 1 || $NUM_DPU -gt 1 ]]
 then
     if [ -f /etc/sonic/database_global.json ]; then
         cp /etc/sonic/database_global.json $REDIS_DIR/sonic-db/database_global.json

files/build_templates/docker_image_ctl.j2

@@ -58,7 +58,7 @@ function updateSyslogConf()
 }
 function ebtables_config()
 {
-    if [ "$DEV" ]; then
+    if [[ "$DEV" && $DATABASE_TYPE != "dpudb" ]]; then
         # Install ebtables filter in namespaces on multi-asic.
         ip netns exec $NET_NS ebtables-restore < /etc/ebtables.filter.cfg
     else
@@ -169,7 +169,7 @@ function postStartAction()
 {%- if docker_container_name == "database" %}
     CHASSISDB_CONF="/usr/share/sonic/device/$PLATFORM/chassisdb.conf"
     [ -f $CHASSISDB_CONF ] && source $CHASSISDB_CONF
-    if [ "$DEV" ]; then
+    if [[ "$DEV" && $DATABASE_TYPE != "dpudb" ]]; then
         # Enable the forwarding on eth0 interface in namespace.
         SYSCTL_NET_CONFIG="/etc/sysctl.d/sysctl-net.conf"
         docker exec -i database$DEV sed -i -e "s/^net.ipv4.conf.eth0.forwarding=0/net.ipv4.conf.eth0.forwarding=1/;
@@ -223,7 +223,7 @@ function postStartAction()
     # databases are not availbale until database container is ready.
     # also chassisdb doesn't support warm/fast reboot, its dump.rdb is deleted
     # at service startup time, nothing need to be done here.
-    if [ "$DATABASE_TYPE" != "chassisdb" ]; then
+    if [[ "$DATABASE_TYPE" != "chassisdb" ]]; then
         # Wait until supervisord and redis starts. This change is needed
         # because now database_config.json is jinja2 templated based
         # and by the time file gets generated if we do redis ping
@@ -503,7 +503,7 @@ start() {
     fi
     {%- endif %}
 
-    if [ -z "$DEV" ]; then
+    if [[ -z "$DEV" || $DATABASE_TYPE == "dpudb" ]]; then
         NET="host"
 
         # For Multi-ASIC platform we have to mount the redis paths for database instances running in different
@@ -528,6 +528,11 @@ start() {
             DB_OPT=$DB_OPT" --env DATABASE_TYPE=$DATABASE_TYPE"
         else
             DB_OPT=$DB_OPT" -v /var/run/redis$DEV:/var/run/redis:rw "
+            DB_OPT=$DB_OPT" --env DATABASE_TYPE=$DATABASE_TYPE "
+            DB_OPT=$DB_OPT" --env NUM_DPU=$NUM_DPU "
+            if [[ "$DEV" ]]; then
+                DB_OPT=$DB_OPT" -v /var/run/redis$DEV:/var/run/redis$DEV:rw "
+            fi
         fi
         {%- endif %}
     else
@@ -557,6 +562,12 @@ start() {
         fi
     fi
 {%- endif %}
+
+    NAMESPACE_ID="$DEV"
+    if [[ $DATABASE_TYPE == "dpudb" ]]; then
+        NAMESPACE_ID=""
+    fi
+
 {%- if sonic_asic_platform == "mellanox" %}
     # TODO: Mellanox will remove the --tmpfs exception after SDK socket path changed in new SDK version
 {%- endif %}
@@ -630,9 +641,10 @@ start() {
 {%- if mount_default_tmpfs|default("n") == "y" %}
         --tmpfs /var/tmp \
 {%- endif %}
-        --env "NAMESPACE_ID"="$DEV" \
+        --env "NAMESPACE_ID"="$NAMESPACE_ID" \
         --env "NAMESPACE_PREFIX"="$NAMESPACE_PREFIX" \
-        --env "NAMESPACE_COUNT"=$NUM_ASIC \
+        --env "NAMESPACE_COUNT"="$NUM_ASIC" \
+        --env "DEV"="$DEV" \
         --env "CONTAINER_NAME"=$DOCKERNAME \
         --name=$DOCKERNAME \
 {%- if docker_container_name == "gbsyncd" %}
@@ -668,7 +680,7 @@ wait() {
 stop() {
     {%- if docker_container_name == "database" %}
     docker stop $DOCKERNAME
-    if [ "$DEV" ]; then
+    if [[ "$DEV" && $DATABASE_TYPE != "dpudb" ]]; then
         ip netns delete "$NET_NS"
     fi
     {%- elif docker_container_name == "teamd" %}
@@ -684,7 +696,7 @@ stop() {
 kill() {
     {%- if docker_container_name == "database" %}
     docker kill $DOCKERNAME
-    if [ "$DEV" ]; then
+    if [[ "$DEV" && $DATABASE_TYPE != "dpudb" ]]; then
         ip netns delete "$NET_NS"
     fi
     {%- else %}
@@ -701,11 +713,16 @@ if [ "$DEV" == "chassisdb" ]; then
       DOCKERNAME=$DOCKERNAME"-chassis"
       unset DEV
 fi
+
+if [[ "$DEV" == *"dpu"* ]]; then
+    DATABASE_TYPE="dpudb"
+fi
+
 {%- endif %}
 NAMESPACE_PREFIX="asic"
 DOCKERNAME=$DOCKERNAME$DEV
 CONTAINER_EXISTS="no"
-if [ "$DEV" ]; then
+if [[ "$DEV" && $DATABASE_TYPE != "dpudb" ]]; then
     NET_NS="$NAMESPACE_PREFIX$DEV" #name of the network namespace
 
     SONIC_CFGGEN="sonic-cfggen -n $NET_NS"

src/sonic-py-common/sonic_py_common/device_info.py

@@ -45,6 +45,9 @@ CHASSIS_INFO_SERIAL_FIELD = 'serial'
 CHASSIS_INFO_MODEL_FIELD = 'model'
 CHASSIS_INFO_REV_FIELD = 'revision'
 
+# DPU constants
+DPU_NAME_PREFIX = "dpu"
+
 # Cacheable Objects
 sonic_ver_info = {}
 hw_info_dict = {}
@@ -841,3 +844,26 @@ def is_frontend_port_present_in_host():
         if not namespace_id:
             return False
     return True
+
+
+def get_num_dpus():
+    # Todo: we should use platform api to get the dpu number
+    # instead of rely on the platform env config.
+    num_dpus = 0
+    platform_env_conf_file_path = get_platform_env_conf_file_path()
+
+    # platform_env.conf file not present for platform
+    if platform_env_conf_file_path is None:
+        return num_dpus
+
+    # Else open the file check for keyword - num_dpu -
+    with open(platform_env_conf_file_path) as platform_env_conf_file:
+        for line in platform_env_conf_file:
+            tokens = line.split('=')
+            if len(tokens) < 2:
+               continue
+            if tokens[0].lower() == 'num_dpu':
+                num_dpus = tokens[1].strip()
+                break
+    return int(num_dpus)
+