# TiDB solutions for avoiding downtime


![left](./student-guide/diagram/28-TiDB-system-unavailable-solution-01-01.png)

------------------------------------------------------------
Unplanned downtime is mainly the result of computer failure or data failure. Planned downtime is mainly due to data changes or system changes:

Failover allows you to limit the crash/recovery time. TiDB Server is stateless, does not affect each other between instances, and requires support from the business side to complete failover.

The BR tool performs database backup and recover.

Rapid data recover can quickly resolve human errors.

TiKV's multi raft mechanism provides high availability at the data replica level.

The etcd embedded in the PD also uses raft to ensure data consistency and availability.

TiDC and TiDB binlog are the replication components of the TiDB database and are the foundation for disaster recover.

With online redefinition, Oracle Database supports many maintenance operations without disrupting database operations or preventing users from updating or accessing data.

The TiDB database supports application patches and database software upgrades on a rolling basis.

The TiDB database continues to expand support for dynamic reconfiguration, enabling it to adapt to changes in demand and hardware without disrupting services.

With online ddl, TiDB databases support many maintenance operations without interrupting database operations or preventing users from updating or accessing data.