nginx重启后，*失败之问题排查记录

问题与排查过程

本地开发环境的服务器，部署了nginx，nginx上对静态的web前端页面进行了http 80端口代理；然后呢，因为一些原因，服务器重启了，重启服务器后，我去把nginx启动起来，但是nginx怎么也代理不了80端口，问题如下：

1. 浏览器访问没响应，本地telnet服务器的80端口是通的，使用wireshark抓包，过程如下：

   

   包3-包7，tcp 三次握手；

   包8，本机给开发服务器发了http请求；

   包9，本机重传；

   包10，服务器返回ack，确认收到请求，但是依然没有请求返回。

   后续一直没有请求返回

2. 服务器端，将nginx.conf进行了精简，精简后，主要配置如下：

   server {
           listen       80;
           server_name  localhost;
           gzip on;
           gzip_http_version 1.1;
           gzip_comp_level 3;
           gzip_types text/plain application/json application/javascript text/css  image/jpeg image/gif image/png application/zip;
   
           access_log  logs/host.access.log;
           
           #这里对web前端静态页面做了代理，路径为/police3-web/scm
           location / {
               root   /police3-web/scm;
               try_files $uri index.html /index.html;
               if ($request_filename ~* .(gif|jpg|jpeg|png|css|js|ico|eot|otf|fon|font|ttf|ttc|woff|woff2)$) {
                 expires   7d;
               }
           }
           location ^~ /scm {
              proxy_set_header X-Real-IP $remote_addr;
              proxy_set_header Host $host;
              proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
                proxy_pass http://localhost:9080;
   
           }
   
           location ~/group([0-9])/M([0-9])([0-9]) {
                   add_header Access-Control-Allow-Origin *;
                   add_header Access-Control-Allow-Methods 'GET, POST, OPTIONS';
                   add_header Access-Control-Allow-Headers 'DNT,X-Mx-ReqToken,Keep-Alive,User-Agent,X-Requested-With,If-Modified-Since,Cache-Control,Content-Type,Authorization';
                   #ngx_fastdfs_module;//因为怀疑和fastdfs有关系，专门注释了
           }
   
       }
   
   

3. 重启nginx后，观察logs/error.log：

   2019/12/06 08:56:53 [alert] 10588#0: worker process 10802 exited on signal 11 (core dumped)
   2019/12/06 08:56:53 [alert] 10588#0: worker process 10802 exited on signal 11 (core dumped)
   2019/12/06 08:56:53 [alert] 10588#0: worker process 10802 exited on signal 11 (core dumped)
   2019/12/06 08:56:53 [notice] 10588#0: start worker process 10806
   2019/12/06 08:56:53 [notice] 10588#0: start worker process 10806
   ngx_http_fastdfs_process_init pid=10806
   2019/12/06 08:56:56 [notice] 10588#0: signal 17 (SIGCHLD) received from 10806
   2019/12/06 08:56:56 [notice] 10588#0: signal 17 (SIGCHLD) received from 10806
   2019/12/06 08:56:56 [alert] 10588#0: worker process 10806 exited on signal 11 (core dumped)
   2019/12/06 08:56:56 [alert] 10588#0: worker process 10806 exited on signal 11 (core dumped)
   2019/12/06 08:56:56 [alert] 10588#0: worker process 10806 exited on signal 11 (core dumped)
   2019/12/06 08:56:56 [notice] 10588#0: start worker process 10808
   2019/12/06 08:56:56 [notice] 10588#0: start worker process 10808
   ngx_http_fastdfs_process_init pid=10808
   2019/12/06 08:56:59 [notice] 10588#0: signal 17 (SIGCHLD) received from 10808
   2019/12/06 08:56:59 [notice] 10588#0: signal 17 (SIGCHLD) received from 10808
   2019/12/06 08:56:59 [alert] 10588#0: worker process 10808 exited on signal 11 (core dumped)
   2019/12/06 08:56:59 [alert] 10588#0: worker process 10808 exited on signal 11 (core dumped)
   2019/12/06 08:56:59 [alert] 10588#0: worker process 10808 exited on signal 11 (core dumped)
   2019/12/06 08:56:59 [notice] 10588#0: start worker process 10812
   2019/12/06 08:56:59 [notice] 10588#0: start worker process 10812
   

   差不多就是循环往复地打这些日志，然后我就去百度了一下，查到了这篇：

   https://blog.csdn.net/hexuan1/article/details/45222867

   里面提到了dmesg命令，我这边也执行了一下，输出如下：

   [55522.992453] nginx[10844]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55526.018405] nginx[10851]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55529.043590] nginx[10853]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55532.070932] nginx[10855]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55535.097444] nginx[10859]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55538.122466] nginx[10861]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55541.148914] nginx[10871]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55544.201955] nginx[10914]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55547.229135] nginx[10921]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55550.254445] nginx[10923]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55553.279543] nginx[10928]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55556.305874] nginx[10932]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55559.331128] nginx[10936]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55562.356655] nginx[10938]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55565.408411] nginx[10943]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55568.434908] nginx[10948]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   [55571.459719] nginx[10950]: segfault at 4 ip 00007f23d017ff84 sp 00007ffdf82783e0 error 4 in libfdfsclient.so[7f23d0179000+18000]
   

   我注意到，这里面提到了：libfdfsclient.so，因为我们的nginx对fastdfs进行了代理（这块同事弄的，暂时没问具体机制，搜了一下，大概如下：https://blog.csdn.net/qq_34301871/article/details/80060235）

   我就想，之前同事好像和我说，要把fastdfs重新启动起来，我一想，可能服务器重启后，fastdfs没启动导致的。

   然后执行了一下命令，重启了fastdfs相关服务：

   /etc/init.d/fdfs_trackerd start
   /etc/init.d/fdfs_storaged start
   

   然后再观察nginx的error.log，居然就没有一直打印上面的错误日志了：

   2019/12/06 08:58:21 [notice] 10588#0: signal 17 (SIGCHLD) received from 10950
   2019/12/06 08:58:21 [notice] 10588#0: signal 17 (SIGCHLD) received from 10950
   2019/12/06 08:58:21 [alert] 10588#0: worker process 10950 exited on signal 11 (core dumped)
   2019/12/06 08:58:21 [alert] 10588#0: worker process 10950 exited on signal 11 (core dumped)
   2019/12/06 08:58:21 [alert] 10588#0: worker process 10950 exited on signal 11 (core dumped)
   2019/12/06 08:58:21 [notice] 10588#0: start worker process 10954
   2019/12/06 08:58:21 [notice] 10588#0: start worker process 10954
   ngx_http_fastdfs_process_init pid=10954
     ##重启fastdfs服务后，这里日志就稳定在下面这一行，没有一直打印错误了：
   [2019-12-06 08:58:23] INFO - fastdfs apache / nginx module v1.21, response_mode=proxy, base_path=/tmp, url_have_group_name=1, group_name=group1, storage_server_port=23000, path_count=1, store_path0=/home/fastdfs/storage, connect_timeout=2, network_timeout=30, tracker_server_count=1, if_alias_prefix=, local_host_ip_count=3, anti_steal_token=0, token_ttl=0s, anti_steal_secret_key length=0, token_check_fail content_type=, token_check_fail buff length=0, load_fdfs_parameters_from_tracker=1, storage_sync_file_max_delay=86400s, use_storage_id=0, storage server id/ip count=0 / 0, flv_support=1, flv_extension=flv
   2019/12/06 09:00:32 [info] 10954#0: *2 client timed out (110: Connection timed out) while waiting for request, client: 10.15.4.46, server: 0.0.0.0:80
   

4. 那，错误排查就到这里了。略坑略坑。

参考

dmesg命令：

https://www.runoob.com/linux/linux-comm-dmesg.html

https://www.cnblogs.com/zhaoxuguang/p/7810651.html