fix(ci): deploy 适配容器化 nginx（pangolin-edge）

nginx 已从主机迁入 pangolin-edge 容器（host 网络模式），其
/etc/nginx/conf.d 由主机 ~/pangolin/edge/conf.d bind-mount。
deploy 改为：jiu.conf 写入该挂载目录 + docker exec pangolin-edge
nginx -t/-s reload，不再 cp 到主机 /etc/nginx 或 systemctl 操作
已废弃的主机 nginx。

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

CHANGELOG.md

@@ -5,6 +5,11 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.0.9] - 2026-06-04
+
+### 修复
+- 适配 nginx 容器化部署：nginx 已迁入 `pangolin-edge` 容器（host 网络），deploy 改为将 jiu.conf 写入 pangolin 挂载目录并 `docker exec pangolin-edge nginx -s reload`，不再操作已废弃的主机 nginx
+
 ## [1.0.8] - 2026-06-03
 
 ### 修复

scripts/ci/deploy.sh

@@ -126,13 +126,12 @@ mkdir -p /opt/jiu/marketing
 rsync -a --delete /tmp/jiu-marketing-new/ /opt/jiu/marketing/
 rm -rf /tmp/jiu-marketing-new
 
-# Update nginx config and reload.
+# Update jiu nginx config in the pangolin-edge reverse proxy.
-# reload-or-restart: reload if nginx is running, start it if it's not — avoids
+# nginx now runs inside the `pangolin-edge` container (host networking), not on
-# `nginx -s reload` failing with "invalid PID number" when nginx isn't running
+# the host. Its /etc/nginx/conf.d is bind-mounted from the host dir below, so we
-# (e.g. after an EC2 reboot where nginx didn't auto-start).
+# write the config there and reload nginx *inside the container*.
-sudo cp /tmp/nginx-jiu.conf /etc/nginx/conf.d/jiu.conf
+cp /tmp/nginx-jiu.conf /home/ec2-user/pangolin/edge/conf.d/jiu.conf
-sudo systemctl enable nginx >/dev/null 2>&1 || true   # auto-start on future reboots
+docker exec pangolin-edge nginx -t && docker exec pangolin-edge nginx -s reload
-sudo nginx -t && sudo systemctl reload-or-restart nginx
 
 echo "Deploy complete!"
 ENDSSH