Cluster Backup
A cluster instance consists of the Mongos, ConfigServer, and Shard nodes, of which the Mongos nodes provide routing capabilities and also provide a unique external application interface for the cluster. The ConfigServer nodes store the metadata of the cluster, which reflects all the state and organization of the cluster. The Shard nodes provide data storage capabilities for the cluster. Therefore, data backup is performed on the ConfigServer and Shard nodes. To minimize the impact on the cluster when backing up data, data backup is performed in Hidden nodes in the ConfigServer and Shard nodes. When a user initiates data backup or the periodical backup time is reached, the instance data is backed up to the remote ZOS from the Hidden node. Therefore, the backup data does not occupy the storage space of the instance.
Replica Set Backup
To minimize the impact of data backup on the database instance service, data backup is performed in Hidden nodes in the replica set instance. When a user initiates data backup or the periodical backup time is reached, the instance data is backed up to the remote ZOS from the Hidden node. Therefore, the backup data does not occupy the storage space of the instance.
Stand-alone Version (Single-Node) Backup
The stand-alone version (single node) contains only 1 node. When a user initiates data backup or the periodical backup time is reached, the instance data is backed up to the remote ZOS from the node. Therefore, the backup data does not occupy the storage space of the instance. However, the backup task consumes the CPU and memory resources of the instance. Therefore, the backup task may fail due to insufficient CPU and memory resources.
Cross-AZ Backup
Instances can be backed up across AZs in multi-AZ resource pools but users cannot select AZs by themselves.