Disaster Recovery

1. Disaster Recovery

Disaster recovery is referred to as disaster recovery and backup. Scheme + = disaster recovery backup scheme disaster recovery solution

DR: the system is deployed in a different city or the available remote regions [], can be performed between each health monitoring and high availability switching occurs when an abnormal stop working area available, the whole system can switch to another application office, does not affect the use of the function. It focused on data synchronization and system continuously available.
Backup: refers to important application data generated by the system or pay would make a copy of the file, increase data security. Focusing on backup and store data.

RTO- recovery time objective: that the system downtime to business services recovery time, RTO is lower the higher the resilience of the system, the cost should be relatively greater.
RPO- recovery point objective: that after a disaster occurs, the system data recovery, recover data corresponding to the point in time called RPO, reflecting the amount of data loss, RPO smaller the business, the less allow lost data. For example, "RPO = 1 day" the day before when using the data recovery means, the data within a day lost.

Back Up
- Operating system backup
- data backup
The amount of backup data
- Full backup: back up all data
- Incremental backup: The incremental backup made since the last backup (the backup based on the last backup)
- Differential Backup: (differential backup once only on the total amount of contrast) in accordance with the last differential backup full backup made
Backup form
- Physical backup: Backup actual data
- Logical backup: backup operation record
The need to stop the service when the backup
- Cold Backup: The data to isolate way to save, without affecting the original data, data recovery slow to address human errors.
- Hot Backup: to build a redundant environment, fast speed of recovery, can not solve human errors.

Active / Passive HA: the cluster comprises only two nodes referred to standby . In this configuration, the system employs primary and backup machines to provide services, the system services only on the master device. When the primary unit fails, the backup service device is activated to replace the service provided by the master device. Typically, CRM software may be used to control such Pacemaker switching between master and slave devices, and a virtual machine to provide IP services.
Active / Active HA: Acronym cluster includes only two nodes double main , into a multi-master (Multi-master) node includes a plurality of time. In this configuration, the system runs the same load on all servers in the cluster. To the database as an example, to update an instance, it will be synchronized to all instances. In this configuration often use load balancing software such as HAProxy to provide virtual IP services.