Backup
CrateDB supports backups with incremental snapshots, ensuring that only new or modified data is stored in subsequent backup snapshots. Snapshots contain the metadata - like users and privileges - and state of the tables in a CrateDB cluster at the time the snapshot was created and can be restored into the cluster at any time.
Snapshots can be stored on the local filesystem, S3 buckets, Azure Blob Storage or any location accessible via the http(s), ftp(s) or file protocols.
Snapshots can be created and restored via standard SQL commands and allow to either take a complete snapshot of metadata and data or be limited to a subset of tables or individual partitions. This is especially helpful, when historical data should be removed from the database, but must be available long-term.
CREATE SNAPSHOT repository_name.snapshot_name { TABLE table_ident [ PARTITION (partition_column = value [, ...])] [, ...] | ALL } [ WITH (snapshot_parameter [= value], [, ...]) ]
Like with the creation of a snapshot, users can decide to restore one of more tables, partitions or metadata, i.e., table and index definitions, views, analyzers, user-defined functions as well as users and privileges, from an existing snapshot into the cluster.