mysql InnoDB的崩溃恢复过程是什么

mysql InnoDB的崩溃恢复过程是什么

1、redo log操作：保证已提交事务影响的最新数据刷到数据页里。

2、undo log操作：保证未提交事务影响的数据页回滚。

3、写缓冲(change buffer)合并。

4、purge操作。

InnoDB的一种垃圾收集机制，使用单独的后台线程周期性处理索引中标记删除的数据。

实例

/* Look for MLOG_CHECKPOINT. */recv_group_scan_log_recs(group, &contiguous_lsn, false);/* The first scan should not have stored or applied any records. */ut_ad(recv_sys->n_addrs == 0);ut_ad(!recv_sys->found_corrupt_fs); if (recv_sys->found_corrupt_log && !srv_force_recovery) {log_mutex_exit();return(DB_ERROR);} if (recv_sys->mlog_checkpoint_lsn == 0) {if (!srv_read_only_mode && group->scanned_lsn != checkpoint_lsn) {ib::error() << "Ignoring the redo log due to missing"" MLOG_CHECKPOINT between the checkpoint "<< checkpoint_lsn << " and the end "<< group->scanned_lsn << ".";if (srv_force_recovery < SRV_FORCE_NO_LOG_REDO) {log_mutex_exit();return(DB_ERROR);}} group->scanned_lsn = checkpoint_lsn;rescan = false;} else {contiguous_lsn = checkpoint_lsn;rescan = recv_group_scan_log_recs(group, &contiguous_lsn, false); if ((recv_sys->found_corrupt_log && !srv_force_recovery) || recv_sys->found_corrupt_fs) {log_mutex_exit();return(DB_ERROR);}} /* NOTE: we always do a 'recovery' at startup, but only ifthere is something wrong we will print a message to theuser about recovery: */ if (checkpoint_lsn != flush_lsn) { if (checkpoint_lsn + SIZE_OF_MLOG_CHECKPOINT < flush_lsn) {ib::warn() << " Are you sure you are using the"" right ib_logfiles to start up the database?"" Log sequence number in the ib_logfiles is "<< checkpoint_lsn << ", less than the"" log sequence number in the first system"" tablespace file header, " << flush_lsn << ".";} if (!recv_needed_recovery) { ib::info() << "The log sequence number " << flush_lsn<< " in the system tablespace does not match"" the log sequence number " << checkpoint_lsn<< " in the ib_logfiles!"; if (srv_read_only_mode) {ib::error() << "Can't initiate database"" recovery, running in read-only-mode.";log_mutex_exit();return(DB_READ_ONLY);} recv_init_crash_recovery();}} log_sys->lsn = recv_sys->recovered_lsn; if (recv_needed_recovery) {err = recv_init_crash_recovery_spaces(); if (err != DB_SUCCESS) {log_mutex_exit();return(err);} if (rescan) {contiguous_lsn = checkpoint_lsn;recv_group_scan_log_recs(group, &contiguous_lsn, true); if ((recv_sys->found_corrupt_log && !srv_force_recovery) || recv_sys->found_corrupt_fs) {log_mutex_exit();return(DB_ERROR);}}} else {ut_ad(!rescan || recv_sys->n_addrs == 0);}登录后复制

版权声明：本文内容由网络用户投稿，版权归原作者所有，本站不拥有其著作权，亦不承担相应法律责任。如果您发现本站中有涉嫌抄袭或描述失实的内容，请联系我们jiasou666@gmail.com 处理，核实后本网站将在24小时内删除侵权内容。