💎一站式轻松地调用各大LLM模型接口,支持GPT4、智谱、星火、月之暗面及文生图 广告
## upstream prematurely 报错如下: >2020/03/10 22:27:31 [crit] 18418#0: *127334 cache file "/www/server/nginx/proxy_cache_dir/d/0c/b14a24698ff6036ae231e21d122070cd" has too long header, client: 111.202.101.156, server: www.dyhjw.com, request: "GET /gold/20200309-55797.html HTTP/1.1", host: "www.dyhjw.com" 解决方法:调小了**keep-alive数目,错误不再出现了** ![](https://img.kancloud.cn/1f/0b/1f0b640bd260bbeffe2b40a766c5bdf2_683x311.png)