diff --git a/backend/services/deploy-mining.sh b/backend/services/deploy-mining.sh
index 923e79d0..61a2054a 100755
--- a/backend/services/deploy-mining.sh
+++ b/backend/services/deploy-mining.sh
@@ -881,16 +881,16 @@ full_reset() {
     fi
 
     echo ""
-    log_step "Step 1/16: Stopping 2.0 services..."
+    log_step "Step 1/18: Stopping 2.0 services..."
     for service in "${MINING_SERVICES[@]}"; do
         service_stop "$service"
     done
 
-    log_step "Step 2/16: Waiting for Kafka consumers to become inactive..."
+    log_step "Step 2/18: Waiting for Kafka consumers to become inactive..."
     log_info "Waiting 15 seconds for consumer group session timeout..."
     sleep 15
 
-    log_step "Step 3/16: Resetting CDC consumer offsets..."
+    log_step "Step 3/18: Resetting CDC consumer offsets..."
     # Reset offsets BEFORE migrations (which may start containers)
     for group in "${CDC_CONSUMER_GROUPS[@]}"; do
         log_info "Resetting consumer group: $group"
@@ -927,17 +927,27 @@ full_reset() {
         fi
     done
 
-    log_step "Step 4/16: Dropping 2.0 databases..."
+    log_step "Step 4/18: Deleting Debezium outbox connectors..."
+    # Delete connectors BEFORE dropping databases to release replication slots
+    local connectors=("auth-outbox-connector" "contribution-outbox-connector" "mining-outbox-connector" "trading-outbox-connector" "mining-wallet-outbox-connector")
+    for connector in "${connectors[@]}"; do
+        log_info "Deleting connector: $connector"
+        curl -s -X DELETE "http://localhost:8084/connectors/$connector" 2>/dev/null || true
+    done
+    log_info "Waiting 5 seconds for connectors to be fully removed..."
+    sleep 5
+
+    log_step "Step 5/18: Dropping 2.0 databases..."
     db_drop
 
-    log_step "Step 5/16: Creating 2.0 databases..."
+    log_step "Step 6/18: Creating 2.0 databases..."
     db_create
 
-    log_step "Step 6/16: Running migrations..."
+    log_step "Step 7/18: Running migrations..."
     db_migrate
 
     # Stop any containers that were started during migration
-    log_step "Step 7/16: Stopping containers and resetting CDC offsets again..."
+    log_step "Step 8/18: Stopping containers and resetting CDC offsets again..."
     log_info "Migration may have started CDC consumers, stopping them now..."
     for service in "${MINING_SERVICES[@]}"; do
         docker compose -f "$COMPOSE_FILE" --env-file "$ENV_FILE" stop "$service" 2>/dev/null || true
@@ -981,22 +991,21 @@ full_reset() {
         fi
     done
 
-    log_step "Step 8/16: Registering Debezium outbox connectors..."
-    # Register outbox connectors for 2.0 service events
-    # These connectors capture events from each service's outbox table and send to Kafka
-    # mining-admin-service consumes these events to aggregate data from all 2.0 services
-    register_outbox_connectors || log_warn "Some connectors may not be registered"
-
-    log_step "Step 9/16: Starting 2.0 services..."
+    log_step "Step 9/18: Starting 2.0 services..."
     for service in "${MINING_SERVICES[@]}"; do
         service_start "$service"
     done
 
-    log_step "Step 10/16: Waiting for services to be ready..."
-    log_info "Waiting 20 seconds for all services to start and sync from 1.0 CDC..."
-    sleep 20
+    log_step "Step 10/18: Waiting for services to be ready and sync from 1.0..."
+    log_info "Waiting 30 seconds for all services to start and sync data from 1.0 CDC..."
+    sleep 30
 
-    log_step "Step 11/16: Publishing legacy users to mining-admin-service..."
+    log_step "Step 11/18: Registering Debezium outbox connectors..."
+    # Register outbox connectors AFTER services are running and have synced data
+    # This ensures outbox_events tables exist and contain data to be captured
+    register_outbox_connectors || log_warn "Some connectors may not be registered"
+
+    log_step "Step 12/18: Publishing legacy users to mining-admin-service..."
     # 调用 auth-service API 发布所有旧用户事件到 outbox
     # 这样 mining-admin-service 才能通过 Debezium 收到用户数据
     local publish_url="http://localhost:3024/api/v2/admin/legacy-users/publish-all"
@@ -1012,7 +1021,11 @@ full_reset() {
         log_info "You may need to manually call: curl -X POST $publish_url"
     fi
 
-    log_step "Step 12/16: Publishing contribution data to mining-admin-service..."
+    log_step "Step 13/18: Waiting for connectors to start capturing..."
+    log_info "Waiting 10 seconds for Debezium connectors to initialize..."
+    sleep 10
+
+    log_step "Step 14/18: Publishing contribution data to mining-admin-service..."
     # 调用 contribution-service API 发布所有算力账户事件到 outbox
     local contrib_publish_url="http://localhost:3020/api/v2/admin/contribution-accounts/publish-all"
     local contrib_result
@@ -1027,7 +1040,7 @@ full_reset() {
         log_info "You may need to manually call: curl -X POST $contrib_publish_url"
     fi
 
-    log_step "Step 13/16: Publishing referral relationships to mining-admin-service..."
+    log_step "Step 15/18: Publishing referral relationships to mining-admin-service..."
     # 调用 contribution-service API 发布所有推荐关系事件到 outbox
     local referral_publish_url="http://localhost:3020/api/v2/admin/referrals/publish-all"
     local referral_result
@@ -1042,7 +1055,7 @@ full_reset() {
         log_info "You may need to manually call: curl -X POST $referral_publish_url"
     fi
 
-    log_step "Step 14/16: Publishing adoption records to mining-admin-service..."
+    log_step "Step 16/18: Publishing adoption records to mining-admin-service..."
     # 调用 contribution-service API 发布所有认种记录事件到 outbox
     local adoption_publish_url="http://localhost:3020/api/v2/admin/adoptions/publish-all"
     local adoption_result
@@ -1063,7 +1076,7 @@ full_reset() {
     # - Calling publish-all again would cause duplicate records in mining-admin-service
     # - See: contribution-calculation.service.ts -> publishContributionRecordEvents()
 
-    log_step "Step 15/16: Publishing network progress to mining-admin-service..."
+    log_step "Step 17/18: Publishing network progress to mining-admin-service..."
     # 调用 contribution-service API 发布全网进度事件到 outbox
     local progress_publish_url="http://localhost:3020/api/v2/admin/network-progress/publish"
     local progress_result
@@ -1076,7 +1089,7 @@ full_reset() {
         log_info "You may need to manually call: curl -X POST $progress_publish_url"
     fi
 
-    log_step "Step 16/16: Waiting for mining-admin-service to sync all data..."
+    log_step "Step 18/18: Waiting for mining-admin-service to sync all data..."
     # 等待 mining-admin-service 消费 outbox 事件
     log_info "Waiting 15 seconds for mining-admin-service to sync all data..."
     sleep 15