ToB企服应用市场:ToB评测及商务社交产业平台

标题: 技术分享 | kill掉mysqld_safe进程会影响mysqld进程？ [打印本页]

---

作者: 立山    时间: 2023-8-14 10:09
标题: 技术分享 | kill掉mysqld_safe进程会影响mysqld进程？
1、背景

公司内部看到一则问题

1、kill -9 mysqld_safe 进程
2、systemd 检测到 mysqld_safe 进程不存在后，重新拉起 mysqld_safe 进程
3、mysqld_safe 进程启动后，发现 mysqld 进程也被重启

期望：启、停 mysqld_safe 进程，不会影响 mysqld 进程
2、systemd 服务启动

2.1、复现问题

1)、查看数据库服务状态

1. [greatsql@greatsql-1 ~]$ sudo systemctl status db-4306
2. ● db-4306.service - db-4306 Server
3.    Loaded: loaded (/usr/lib/systemd/system/db-4306.service; disabled; vendor preset: disabled)
4.    Active: active (running) since Wed 2023-07-19 11:15:18 CST; 6h ago
5. Main PID: 14917 (mysqld_safe)
6.    CGroup: /system.slice/db-4306.service
7.            ├─14917 /bin/sh /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
8.            └─16340 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data -...
9. Jul 19 11:15:18 greatsql-1 systemd[1]: Started db-4306 Server.
10. Jul 19 11:15:19 greatsql-1 mysqld_safe[14917]: mysqld_safe Adding '/greatsql/svr/GreatSQL-8.0.32-24-Linux-glibc2.17-x86_64/lib/mysql/libjemalloc.so.1' t...or mysqldJul 19 11:15:19 greatsql-1 mysqld_safe[14917]: 2023-07-19T03:15:19.907338Z mysqld_safe Logging to '/greatsql/logs/error4306.log'.
11. Jul 19 11:15:19 greatsql-1 mysqld_safe[14917]: 2023-07-19T03:15:19.953728Z mysqld_safe Starting mysqld daemon with databases from /greatsql/dbdata/data4306/data
12. Hint: Some lines were ellipsized, use -l to show in full.

复制代码

2)、kill -9 mysqld_safe 进程，并再次查看数据库服务状态

1. [greatsql@greatsql-1 ~]$ kill -9 14917
2. [greatsql@greatsql-1 ~]$ sudo systemctl status db-4306
3. ● db-4306.service - db-4306 Server
4.    Loaded: loaded (/usr/lib/systemd/system/db-4306.service; disabled; vendor preset: disabled)
5.    Active: active (running) since Wed 2023-07-19 18:00:33 CST; 43s ago
6. Main PID: 15195 (mysqld_safe)
7.     Tasks: 50
8.    CGroup: /system.slice/db-4306.service
9.            ├─15195 /bin/sh /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
10.            └─16613 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data -...
11. Jul 19 18:00:33 greatsql-1 systemd[1]: Started db-4306 Server.
12. Jul 19 18:00:34 greatsql-1 mysqld_safe[15195]: mysqld_safe Adding '/greatsql/svr/GreatSQL-8.0.32-24-Linux-glibc2.17-x86_64/lib/mysql/libjemalloc.so.1' t...or mysqldJul 19 18:00:34 greatsql-1 mysqld_safe[15195]: 2023-07-19T10:00:34.640240Z mysqld_safe Logging to '/greatsql/logs/error4306.log'.
13. Jul 19 18:00:34 greatsql-1 mysqld_safe[15195]: 2023-07-19T10:00:34.679333Z mysqld_safe Starting mysqld daemon with databases from /greatsql/dbdata/data4306/data
14. Hint: Some lines were ellipsized, use -l to show in full.

复制代码

确实在 kill -9 mysqld_safe 后，重新拉起了 mysqld_safe 和 mysqld 进程(它们的 PID 和之前不一样)
3)、查看数据库错误日志

1. 2023-07-19T18:00:31.933020+08:00 0 [System] [MY-013172] [Server] Received SHUTDOWN from user <via user signal>. Shutting down mysqld (Version: 8.0.32-24).

复制代码

4)、查看 service 文件

1. [greatsql@greatsql-1 ~]$ cat /usr/lib/systemd/system/db-4306.service
2. [Unit]
3. Description=db-4306 Server
4. After=network.target
6. [Install]
7. WantedBy=multi-user.target
9. [Service]
10. User=greatsql
11. Group=greatsql
13. Type=simple
14. ExecStart=/greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
15. Restart=on-failure
16. LimitNOFILE=1024000
17. LimitNPROC=1024000
18. TimeoutStopSec=15
19. PrivateTmp=false

复制代码

2.2、分析原因

1)、查看进程信息

1. [greatsql@greatsql-1 ~]$ ps axj |head -1;ps axj |grep 4306 |grep -v grep
2. PPID   PID  PGID   SID TTY      TPGID STAT   UID   TIME COMMAND
3.     1 15195 15195 15195 ?           -1 Ss     986   0:00 /bin/sh /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
4. 15195 16613 15195 15195 ?           -1 Sl     986   0:06 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data --plugin-dir=/greatsql/svr/greatsql/lib/plugin --log-error=/greatsql/logs/error4306.log --open-files-limit=65535 --pid-file=/greatsql/dbdata/data4306/data/mysql.pid --socket=/greatsql/dbdata/data4306/data/mysql.sock --port=4306

复制代码

mysqld_safe：PID(进程 ID)=PGID(进程组 ID)=SID(会话 ID)，说明它是会话首进程，也是该进程组的组长
mysqld_safe 的 PID(进程 ID)=mysqld 的 PPID(父进程 ID)，说明 mysqld_safe 是 mysqld 的父进程
kill -9 mysqld_safe (会话首进程)，会向该进程组的每一个进程发送 SIGKILL，导致组中的进程被中止
2)、整体流程

• mysqld_safe 是会话首进程，kill -9 mysqld_safe，导致组中所有进程被 kill
  
• systemd 检测到 mysqld_safe 异常退出，Restart=on-failure 触发重新拉起 mysqld_safe
  
• mysqld_safe 拉起子进程 mysqld
  

3、mysqld_safe 命令启动

1)、使用 mysqld_safe 启动数据库

1. [greatsql@greatsql-1 ~]$ /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf &
2. [1] 18229
3. [greatsql@greatsql-1 ~]$  mysqld_safe Adding '/greatsql/svr/GreatSQL-8.0.32-24-Linux-glibc2.17-x86_64/lib/mysql/libjemalloc.so.1' to LD_PRELOAD for mysqld
4. 2023-07-19T14:20:19.135297Z mysqld_safe Logging to '/greatsql/logs/error4306.log'.
5. 2023-07-19T14:20:19.173594Z mysqld_safe Starting mysqld daemon with databases from /greatsql/dbdata/data4306/data

复制代码

2)、查看进程信息

1. [greatsql@greatsql-1 ~]$ ps axj |head -1;ps axj |grep 4306 |grep -v grep
2. PPID   PID  PGID   SID TTY      TPGID STAT   UID   TIME COMMAND
3. 17360 18229 18229 17206 pts/7    17360 S      986   0:00 /bin/sh /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
4. 18229 19658 18229 17206 pts/7    17360 Sl     986   0:02 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data --plugin-dir=/greatsql/svr/greatsql/lib/plugin --log-error=/greatsql/logs/error4306.log --open-files-limit=65535 --pid-file=/greatsql/dbdata/data4306/data/mysql.pid --socket=/greatsql/dbdata/data4306/data/mysql.sock --port=4306

复制代码

PID≠SID，不是会话首进程
PGID≠TPGID，是后台进程组
mysqld_safe 的 PID=mysqld 的 PPID，说明 mysqld_safe 是 mysqld 的父进程
3)、kill -9 mysqld_safe 进程，并再次查看进程信息

1. [greatsql@greatsql-1 ~]$ kill -9 18229
2. [greatsql@greatsql-1 ~]$ ps axj |head -1;ps axj |grep 4306 |grep -v grep
3. PPID   PID  PGID   SID TTY      TPGID STAT   UID   TIME COMMAND
4.     1 19658 18229 17206 pts/7    17360 Sl     986   0:07 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data --plugin-dir=/greatsql/svr/greatsql/lib/plugin --log-error=/greatsql/logs/error4306.log --open-files-limit=65535 --pid-file=/greatsql/dbdata/data4306/data/mysql.pid --socket=/greatsql/dbdata/data4306/data/mysql.sock --port=4306

复制代码

kill -9 mysqld_safe (非会话首进程)，不影响同组的进程(mysqld)，此时 init 进程会自动领养 mysqld 进程
4)、重新启动 mysqld_safe 进程，并再次查看进程信息

1. [greatsql@greatsql-1 ~]$ /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf &
2. [1] 31401
3. [greatsql@greatsql-1 ~]$  mysqld_safe Adding '/greatsql/svr/GreatSQL-8.0.32-24-Linux-glibc2.17-x86_64/lib/mysql/libjemalloc.so.1' to LD_PRELOAD for mysqld
4. 2023-07-19T14:38:42.429733Z mysqld_safe Logging to '/greatsql/logs/error4306.log'.
5. 2023-07-19T14:38:42.493870Z mysqld_safe A mysqld process already exists
7. [1]+  Exit 1                  /greatsql/svr/greatsql/bin/mysqld_safe --defaults-file=/greatsql/conf/greatsql4306.cnf
9. [greatsql@greatsql-1 ~]$ ps axj |head -1;ps axj |grep 4306 |grep -v grep
10. PPID   PID  PGID   SID TTY      TPGID STAT   UID   TIME COMMAND
11.     1 19658 18229 17206 pts/7    17360 Sl     986   0:09 /greatsql/svr/greatsql/bin/mysqld --defaults-file=/greatsql/conf/greatsql4306.cnf --basedir=/greatsql/svr/greatsql --datadir=/greatsql/dbdata/data4306/data --plugin-dir=/greatsql/svr/greatsql/lib/plugin --log-error=/greatsql/logs/error4306.log --open-files-limit=65535 --pid-file=/greatsql/dbdata/data4306/data/mysql.pid --socket=/greatsql/dbdata/data4306/data/mysql.sock --port=4306

复制代码

mysqld_safe 退出，原因是已存在 mysqld 进程
4、总结

• mysqld_safe 进程和 mysqld 进程是父子进程关系
  
• systemd 服务启动，kill 父进程(会话首进程)，会导致子进程也退出
  
• mysqld_safe 命令启动，kill 父进程(非会话首进程)，不会导致子进程退出；重新启动父进程，报错子进程已存在
  
• 可以使用mysqld_safe命令启动 + 改造mysqld_safe脚本，实现【启、停mysqld_safe进程，不会影响mysqld进程】。此时不要混合使用systemd启动数据库，需要维护这个特殊的mysqld_safe
  

Enjoy GreatSQL
免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！

---

欢迎光临 ToB企服应用市场:ToB评测及商务社交产业平台 (https://dis.qidao123.com/)

Powered by Discuz! X3.4