refactor(k01): flatten directory layout and split deployment into five scripts

deploy/k01/README.md

@@ -2,7 +2,7 @@
 
 该目录是 juwan-backend 所有 k3s 节点的初始化配置。公网入口由 center 的 Caddy 接管——`/wt/*` 走 UDP 直达 chat-api，其余路径反代到 envoy-gateway NodePort 30080。
 
-第一台机器按以下步骤初始化为 k3s server；后续加入的 k02、k03 只运行 `install-k3s.sh agent`，其他步骤（k8s Secret、CR、业务 yaml）在 server 上 apply 一次即可。
+第一台机器按以下步骤初始化为 k3s server；后续加入的 k02、k03 只运行 `install.sh agent`，其他步骤在 server 上执行一次即可。
 
 ## 前置条件
 
@@ -12,29 +12,18 @@
 - `/root/registry-password` 文件存放 zot admin 密码（`chmod 600`）
 - `.env` 已按 `.env.example` 填好（zot admin 密码、Brevo SMTP、Garage S3 凭据）
 
-如果还没 `.env`：先 `cp .env.example .env && nano .env`，再跑 `install-k3s.sh`。
+如果还没 `.env`：先 `cp .env.example .env && nano .env`，再跑 `secrets.sh`。
 
 ## k3s server 初始化
 
 ```bash
 cd /root/juwan-backend/deploy/k01
 
-# 装 k3s + Helm + 四个 Operator，并自动调用 secrets.sh 生成全部 k8s Secret
-bash install-k3s.sh
-
-# 应用基础设施层（Operator CR）
-kubectl apply -f 01-infra/postgres.yaml
-kubectl apply -f 01-infra/redis.yaml
-kubectl apply -f 01-infra/mongo.yaml
-kubectl apply -f 01-infra/kafka.yaml
-kubectl apply -f 01-infra/ratelimit.yaml
-kubectl apply -f 01-infra/envoy.yaml
-
-# PostgreSQL Cluster 全部 Ready 后写入 schema 与 fixture
-bash 01-infra/load-schema.sh
-
-# 启动业务服务
-kubectl apply -f 02-service/
+bash install.sh          # k3s + Helm + 四个 Operator
+bash secrets.sh          # 生成所有 k8s Secret
+bash apply-infra.sh      # 数据层 + envoy + ratelimit，分批等待 Ready
+bash apply-schema.sh     # 向 CNPG 写入 schema 与 fixture
+bash apply-services.sh   # 启动业务 Deployment
 
 kubectl -n juwan get pods -w
 ```
@@ -43,7 +32,7 @@ kubectl -n juwan get pods -w
 
 控制面是 k3s server，跑着 CNPG / Strimzi / Redis / MongoDB 四个 Operator 管理有状态服务。
 
-数据层 11 个 per-domain PostgreSQL Cluster + 12 个 RedisReplication + 1 个 MongoDBCommunity（chat）+ Strimzi KRaft Kafka（1 broker）。
+数据层 11 个 per-domain PostgreSQL Cluster + 12 个 RedisReplication + 1 个 MongoDBCommunity + Strimzi KRaft Kafka。
 
 业务层 27 个 Go 服务镜像指向 `registry.juwan.xhttp.zip/juwan/<name>:latest`，每个 domain 一套 rpc + api，外加 snowflake、authz-adapter、email-mq 和 frontend。所有 Deployment 带 `imagePullSecrets: registry-creds`，containerd 的 `registries.yaml` 配了 zot admin 凭据。
 
@@ -55,7 +44,7 @@ chat-api 的 WebTransport 走 UDP 8443 hostPort，center Caddy 的 PR 7669 fork
 
 `secrets.sh` 生成随机密码写入 `secrets/` 目录，同时 `kubectl create secret` 到 `juwan` namespace。需要手动填的是 `.env` 里的 zot admin 密码、Brevo SMTP key 和 Garage S3 access key。
 
-CNPG 每个 Cluster Ready 后自动生成 `<cluster>-app` Secret（username/password/dbname/host/port），业务 pod 的 env 直接从这些 Secret 取值。
+CNPG 每个 Cluster Ready 后自动生成 `<cluster>-app` Secret（username/password/dbname/host/port），业务 pod 的 env 由这些 Secret 提供。
 
 ## 加节点
 
@@ -71,7 +60,7 @@ cat /var/lib/rancher/k3s/server/node-token
 cd /root/juwan-backend/deploy/k01
 echo "<zot-admin-password>" > /root/registry-password && chmod 600 /root/registry-password
 
-K3S_URL=https://<server-ip>:6443 K3S_TOKEN=<token> bash install-k3s.sh agent
+K3S_URL=https://<server-ip>:6443 K3S_TOKEN=<token> bash install.sh agent
 ```
 
 ## 日常操作

deploy/k01/apply-infra.sh

@@ -0,0 +1,29 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+INFRA_DIR="$(cd "$(dirname "$0")/infra" && pwd)"
+export KUBECONFIG="${KUBECONFIG:-/etc/rancher/k3s/k3s.yaml}"
+
+echo envoy + ratelimit
+kubectl apply -f "${INFRA_DIR}/envoy.yaml"
+kubectl apply -f "${INFRA_DIR}/ratelimit.yaml"
+kubectl -n juwan wait --for=condition=Ready pod -l app=envoy-gateway --timeout=120s || true
+kubectl -n juwan wait --for=condition=Ready pod -l "app in (ratelimit,rl-redis)" --timeout=120s || true
+
+echo redis
+kubectl apply -f "${INFRA_DIR}/redis.yaml"
+kubectl -n juwan wait --for=jsonpath='{.status.readyReplicas}'=1 redisreplication --all --timeout=600s || true
+
+echo postgres
+kubectl apply -f "${INFRA_DIR}/postgres.yaml"
+kubectl -n juwan wait --for=condition=Ready cluster --all --timeout=900s || true
+
+echo mongo
+kubectl apply -f "${INFRA_DIR}/mongo.yaml"
+kubectl -n juwan wait --for=jsonpath='{.status.phase}'=Running mongodbcommunity/chat-mongodb --timeout=600s || true
+
+echo kafka
+kubectl apply -f "${INFRA_DIR}/kafka.yaml"
+kubectl -n kafka wait --for=condition=Ready kafka/juwan-kafka --timeout=900s || true
+
+kubectl get pods -A

deploy/k01/apply-schema.sh

@@ -8,19 +8,12 @@ FIXTURE_DIR="$REPO_ROOT/deploy/dev/fixture"
 
 export KUBECONFIG=/etc/rancher/k3s/k3s.yaml
 
-declare -A SCHEMA_MAP=(
-	[user-db]=users
-	[player-db]=player
-	[game-db]=game
-	[shop-db]=shop
-	[order-db]=order
-	[wallet-db]=wallet
-	[community-db]=community
-	[review-db]=review
-	[dispute-db]=dispute
-	[notification-db]=notification
-	[search-db]=search
-)
+domain_dir() {
+	case "$1" in
+		user) echo users ;;
+		*) echo "$1" ;;
+	esac
+}
 
 psql_exec() {
 	local cluster="$1" sql="$2"
@@ -34,25 +27,27 @@ psql_file() {
 		-v ON_ERROR_STOP=1 -U app -d app < "$file"
 }
 
-for cluster in "${!SCHEMA_MAP[@]}"; do
-	domain="${SCHEMA_MAP[$cluster]}"
-	echo ">>> $cluster ($domain)"
+clusters=()
+while IFS= read -r name; do
+	clusters+=("$name")
+done < <(kubectl -n juwan get cluster -o jsonpath='{.items[*].metadata.name}' | tr ' ' '\n')
 
+for cluster in "${clusters[@]}"; do
+	domain="${cluster%-db}"
+	dir="$(domain_dir "$domain")"
+	echo "$cluster"
 	kubectl -n juwan wait --for=condition=Ready "cluster.postgresql.cnpg.io/${cluster}" --timeout=300s
-
 	psql_file "$cluster" "$SQL_DIR/common/update_updated_at_column.sql"
-
-	for f in "$SQL_DIR/$domain"/*.sql; do
+	for f in "$SQL_DIR/$dir"/*.sql; do
 		[ -f "$f" ] || continue
-		echo "    schema: $(basename "$f")"
+		echo "  $(basename "$f")"
 		psql_file "$cluster" "$f"
 	done
-
-	if [ -f "$FIXTURE_DIR/$domain.sql" ]; then
-		echo "    fixture: $domain.sql"
-		psql_file "$cluster" "$FIXTURE_DIR/$domain.sql"
+	if [ -f "$FIXTURE_DIR/$dir.sql" ]; then
+		echo "  $dir.sql"
+		psql_file "$cluster" "$FIXTURE_DIR/$dir.sql"
 	fi
 done
 
 echo
-echo "schema + fixture loaded into 11 CNPG clusters"
+echo "schema + fixture loaded, ${#clusters[@]} clusters"

deploy/k01/apply-services.sh

@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SVC_DIR="$(cd "$(dirname "$0")/services" && pwd)"
+export KUBECONFIG="${KUBECONFIG:-/etc/rancher/k3s/k3s.yaml}"
+
+apply_wait() {
+	for f in "$@"; do
+		echo "${f%.yaml}"
+		kubectl apply -f "${SVC_DIR}/${f}"
+	done
+	kubectl -n juwan wait --for=condition=Available deploy --all --timeout=600s || true
+}
+
+cd "$SVC_DIR"
+
+apply_wait snowflake.yaml authz-adapter.yaml
+
+domain_files=()
+for f in *.yaml; do
+	case "$f" in
+		snowflake.yaml|authz-adapter.yaml|chat.yaml|email.yaml|frontend.yaml) ;;
+		*) domain_files+=("$f") ;;
+	esac
+done
+apply_wait "${domain_files[@]}"
+
+apply_wait chat.yaml email.yaml frontend.yaml
+
+kubectl get pods -n juwan

deploy/k01/install.sh

@@ -21,13 +21,6 @@ fi
 
 K01_DIR="$(cd "$(dirname "$0")" && pwd)"
 
-if [ "$MODE" = "server" ] && [ ! -f "${K01_DIR}/.env" ]; then
-	echo ".env not found in ${K01_DIR}" >&2
-	echo "  cp ${K01_DIR}/.env.example ${K01_DIR}/.env" >&2
-	echo "  nano ${K01_DIR}/.env" >&2
-	exit 1
-fi
-
 write_registries() {
 	mkdir -p /etc/rancher/k3s
 	cat > /etc/rancher/k3s/registries.yaml <<EOF
@@ -82,7 +75,7 @@ systemctl restart k3s
 export KUBECONFIG=/etc/rancher/k3s/k3s.yaml
 until kubectl get nodes >/dev/null 2>&1; do sleep 2; done
 
-kubectl apply -f "${K01_DIR}/00-base/"
+kubectl apply -f "${K01_DIR}/base/"
 
 kubectl apply --server-side --force-conflicts -f \
 	"https://github.com/cloudnative-pg/cloudnative-pg/releases/download/v${CNPG_VERSION}/cnpg-${CNPG_VERSION}.yaml"
@@ -121,8 +114,6 @@ kubectl -n kafka rollout status deploy/strimzi-cluster-operator --timeout=300s
 kubectl -n redis-operator rollout status deploy/redis-operator --timeout=300s
 kubectl -n mongodb-operator rollout status deploy/mongodb-kubernetes-operator --timeout=300s
 
-bash "${K01_DIR}/secrets.sh"
-
 echo
 echo "k3s server + 4 operators ready"
 echo "node token: $(cat /var/lib/rancher/k3s/server/node-token)"

deploy/k01/secrets.sh

@@ -69,7 +69,11 @@ kubectl -n juwan create secret tls chat-wt-tls \
 	--key="${DEV_CERTS}/tls.key" \
 	--dry-run=client -o yaml | kubectl apply -f -
 
-DOMAINS=(user player game shop order wallet community review dispute notification search chat)
+DOMAINS=()
+while IFS= read -r name; do
+	DOMAINS+=("${name%-redis}")
+done < <(grep -E '^  name: [a-z-]+-redis$' "$K01_DIR/infra/redis.yaml" | awk '{print $2}')
+
 for d in "${DOMAINS[@]}"; do
 	pwd_val="$(openssl rand -hex 16)"
 	write_secret "redis-${d}-password" "$pwd_val"