首页 > 数据库 > MySQL > 正文

MySQL 5.7复制延迟之sync_relay_log

2024-07-24 12:36:18
字体:
来源:转载
供稿:网友
        MySQL 5.7复制延迟之sync_relay_log:

  一、描述
  MySQL 5.7版本主从复制,批量时候显示延迟上万秒。
 
  二、现象
 
  1、io使用率高
  #iostat -dxm 1 1000
  Device:         rrqm/s   wrqm/s     r/s     w/s    rMB/s    wMB/s avgrq-sz avgqu-sz   await  svctm  %util
  scd0              0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  vda               0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  vdb               0.00    96.00    0.00 2596.00     0.00     8.54     6.74     1.33    0.51   0.37  95.30
  vdc               0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  vdd               0.00     0.00    0.00   11.00     0.00     0.06    11.64     0.00    0.09   0.09   0.10
  vde               0.00     0.00    0.00    7.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  vdf               0.00     0.00    0.00  511.00     0.00     0.00     0.00     0.05    0.09   0.09   4.60
  vdg               0.00     0.00    0.00  511.00     0.00     0.00     0.00     0.05    0.09   0.09   4.80
  dm-0              0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  dm-1              0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  dm-2              0.00     0.00    0.00   34.00     0.00     0.23    13.65     0.02    0.59   0.38   1.30
  dm-3              0.00     0.00    0.00 2144.00     0.00     8.38     8.00     1.40    0.65   0.45  97.20
  dm-4              0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
  dm-5              0.00     0.00    0.00    0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
 
  2、dm3是relay log 和binlog分区
  $ ls -l /dev/mapper
  total 0
  lrwxrwxrwx 1 root root      7 Jul 23 23:20 backup-backup -> ../dm-0
  crw-rw---- 1 root root 10, 58 Jul 23 23:20 control
  lrwxrwxrwx 1 root root      7 Jul 23 23:20 VG00-lv_root -> ../dm-4
  lrwxrwxrwx 1 root root      7 Jul 23 23:20 zxmysql-zxdba -> ../dm-1
  lrwxrwxrwx 1 root root      7 Jul 23 23:20 zxmysql-zxlog -> ../dm-3
 
  3、slave状态
  mysql> show slave status /G;
  *************************** 1. row ***************************
                 Slave_IO_State: Queueing master event to the relay log
                  略.........................................
                  Connect_Retry: 60
                Master_Log_File: mysql-bin.011494
            Read_Master_Log_Pos: 21037034
                 Relay_Log_File: relay-log.001904
                  Relay_Log_Pos: 3154097
          Relay_Master_Log_File: mysql-bin.011494
               Slave_IO_Running: Yes
              Slave_SQL_Running: Yes
                Replicate_Do_DB:
            Replicate_Ignore_DB:
             Replicate_Do_Table:
         Replicate_Ignore_Table:
        Replicate_Wild_Do_Table:
    Replicate_Wild_Ignore_Table:
                     Last_Errno: 0
                     Last_Error:
                   Skip_Counter: 0
            Exec_Master_Log_Pos: 3153884
                Relay_Log_Space: 21037535
                Until_Condition: None
                 Until_Log_File:
                  Until_Log_Pos: 0
             Master_SSL_Allowed: No
             Master_SSL_CA_File:
             Master_SSL_CA_Path:
                Master_SSL_Cert:
              Master_SSL_Cipher:
                 Master_SSL_Key:
          Seconds_Behind_Master: 471
  Master_SSL_Verify_Server_Cert: No
                  Last_IO_Errno: 0
                  Last_IO_Error:
                 Last_SQL_Errno: 0
                 Last_SQL_Error:
    Replicate_Ignore_Server_Ids:
               Master_Server_Id: 400011
                    Master_UUID: 0f8507ea-6da1-11e8-8646-005056873c4a
               Master_Info_File: mysql.slave_master_info
                      SQL_Delay: 0
            SQL_Remaining_Delay: NULL
        Slave_SQL_Running_State: Reading event from the relay log
             Master_Retry_Count: 86400
                    Master_Bind:
        Last_IO_Error_Timestamp:
       Last_SQL_Error_Timestamp:
                 Master_SSL_Crl:
             Master_SSL_Crlpath:
             Retrieved_Gtid_Set: 0f8507ea-6da1-11e8-8646-005056873c4a:14137114-19288497
              Executed_Gtid_Set: 0f8507ea-6da1-11e8-8646-005056873c4a:1-19288446
                  Auto_Position: 1
           Replicate_Rewrite_DB:
                   Channel_Name:
             Master_TLS_Version:
  1 row in set (0.01 sec)
 
  ERROR:
  No query specified
  三、分析
  通过以上现象发现备库io使用率过高,超过90%。io过高的磁盘为日志盘,存放relay log和binlog。io thead一致在写relay log,调用fdatasync写磁盘。这里涉及到一个参数sync_relay_log,默认值为10000,查看当前系统参数值为1.
 
  四、解决方案
  优化io thread线程和sql thread线程。sync_relay_log使用默认值,使用mts优化sql thread。
 
  stop slave;
  set global slave_parallel_type=logical_clock;
  set global slave_parallel_workers=8;
  set global sync_master_info=10000;
  set global sync_relay_log=10000;
  set global sync_relay_log_info=10000;
  start slave;

(编辑:武林网)

发表评论 共有条评论
用户名: 密码:
验证码: 匿名发表