How does innobackupex works

By | 2013-11-11

innobackupex工作流程

innobackupex封装了xtrabackup(InnoDB备份)工具,大致的备份流程如下:

|创建备份目录|  ---> |xtrabackup备份InnoDB| --> |全局锁| --> |备份非InnoDB表| --> |释放锁|

xtrabackup 备份 InnoDB 过程中,InnoDB 相关变化如下显示:

备份过程
上述部分为拷贝文件阶段,文件的变化及拷贝文件工作流程见下图:
拷文件流程
在备份InnoDB的过程中,记录的变更保存于xtrabackup_logfile文件,所以在prepare(–apply-log)的时候,重放该部分数据到表空间即可,如下:
重放过程
以上就完成了在拷贝文件完成后的那个时间点的全量备份;

FAQ:

xtrabackup备份开始时,suspend到底有什么作用?

130716 12:14:08  innobackupex: Starting ibbackup with command: xtrabackup_55  --defaults-file="/web/mysql/node3306/my.node.cnf"  --defaults-group="mysqld" --backup --suspend-at-end --target-dir=/web/xtrabackup/backups/node3306/2013-07-16_12-14-08 --tmpdir=/dev/shm
innobackupex: Waiting for ibbackup (pid=16924) to suspend
innobackupex: Suspend file '/web/xtrabackup/backups/node3306/2013-07-16_12-14-08/xtrabackup_suspended_2'

如下备份信息,为什么要有一个suspend的过程,从innobackupex代码来看,suspend起到了进程同步的作用,即先用xtrabackup备份Innodb,通过suspend作用来检测xtrabackup是否备份正常,是则全局锁,备份其他引擎文件,否则Die失败;

sub backup {
    my $orig_datadir = get_option(\%config, $option_defaults_group, 'datadir');

    # check that we can connect to the database. This done by
    # connecting, issuing a query, and closing the connection.
    mysql_open();
    mysql_close();

    # start ibbackup as a child process
    start_ibbackup();                        -----   开始innodb备份

    # wait for ibbackup to suspend itself
    if (!$option_remote_host) {
        wait_for_ibbackup_suspend();         -----   等待innodb备份成功,否则Die失败,成功则进行到下一步其他引擎的备份;
    }

    # connect to database
    mysql_open();

    if ($option_safe_slave_backup) {
      wait_for_safe_slave();
    }

    # flush tables with read lock
    if (!$option_no_lock) {
        # make a prep copy before locking tables, if using rsync
        backup_files(1);

        # flush tables with read lock
        mysql_lockall();
    }
    ....
}
......
sub wait_for_ibbackup_suspend {
    print STDERR "$prefix Waiting for ibbackup (pid=$ibbackup_pid) to suspend\n";
    print STDERR "$prefix Suspend file '$suspend_file'\n\n";
    for (;;) {
        sleep 2;
        last if -e $suspend_file;           -----检测是否存在suspend文件, 是则跳到最后一次循环;

        # check that ibbackup child process is still alive
        if ($ibbackup_pid == waitpid($ibbackup_pid, &WNOHANG)) {    ---- waitpid检测start_ibbackup()是否完成,异常则Die失败退出;
            $ibbackup_pid = '';
            Die "ibbackup child process has died";
        }
    }
    $now = current_time();
    open XTRABACKUP_PID, "> $option_tmpdir/$xtrabackup_pid_file";
    print XTRABACKUP_PID $ibbackup_pid;
    close XTRABACKUP_PID;
    print STDERR "\n$now  $prefix Continuing after ibbackup has suspended\n";
}

sub resume_ibbackup {
    print STDERR "$prefix Resuming ibbackup\n\n";         
    unlink $suspend_file || Die "Failed to delete '$suspend_file': $!";                -----删除suspend在repare这步;

    # wait for ibbackup to finish
    waitpid($ibbackup_pid, 0);
    unlink "$option_tmpdir/$xtrabackup_pid_file"; 
    $ibbackup_pid = '';
    return $CHILD_ERROR >> 8;
}

2 thoughts on “How does innobackupex works

  1. Pingback: innobackupex restore | cz blog site.

  2. Pingback: 使用Xtrabackup备份远端主机的MySQL实例 | cz

Leave a Reply

Your email address will not be published. Required fields are marked *