BACKUP

On this page

CockroachDB's BACKUP statement allows you to create full or incremental backups of your cluster's schema and data that are consistent as of a given timestamp. Backups can be with or without revision history.

Because CockroachDB is designed with high fault tolerance, these backups are designed primarily for disaster recovery (i.e., if your cluster loses a majority of its nodes) through RESTORE. Isolated issues (such as small-scale node outages) do not require any intervention.

Functional details

Backup targets

You can backup entire tables (which automatically includes their indexes) or views. Backing up a database simply backs up all of its tables and views.

Object dependencies

Dependent objects must be backed up at the same time as the objects they depend on.

Users and privileges

Every backup you create includes system.users, which stores your users and their passwords. To restore your users, you must use this procedure.

Restored tables inherit privilege grants from the target database; they do not preserve privilege grants from the backed up table because the restoring cluster may have different users.

Table-level privileges must be granted to users after the restore is complete.

Backup types

CockroachDB offers two types of backups: full and incremental.

Full backups

Full backups contain an unreplicated copy of your data and can always be used to restore your cluster. These files are roughly the size of your data and require greater resources to produce than incremental backups. You can take full backups as of a given timestamp and (optionally) include the available revision history.

Incremental backups

Incremental backups are smaller and faster to produce than full backups because they contain only the data that has changed since a base set of backups you specify (which must include one full backup, and can include many incremental backups). You can take incremental backups either as of a given timestamp or with full revision history.

Note the following restrictions:

• Incremental backups can only be created within the garbage collection period of the base backup's most recent timestamp. This is because incremental backups are created by finding which data has been created or modified since the most recent timestamp in the base backup––that timestamp data, though, is deleted by the garbage collection process.

  You can configure garbage collection periods using the ttlseconds replication zone setting.

• It is not possible to create an incremental backup if one or more tables were created, dropped, or truncated after the full backup. In this case, you must create a new full backup.

Backups with revision history

You can create full or incremental backups with revision history:

• Taking full backups with revision history allows you to back up every change made within the garbage collection period leading up to and including the given timestamp.
• Taking incremental backups with revision history allows you to back up every change made since the last backup and within the garbage collection period leading up to and including the given timestamp. You can take incremental backups with revision history even when your previous full or incremental backups were taken without revision history.

You can configure garbage collection periods using the ttlseconds replication zone setting. Taking backups with revision history allows for point-in-time restores within the revision history.

Performance

The BACKUP process minimizes its impact to the cluster's performance by distributing work to all nodes. Each node backs up only a specific subset of the data it stores (those for which it serves writes; more details about this architectural concept forthcoming), with no two nodes backing up the same data.

For best performance, we also recommend always starting backups with a specific timestamp at least 10 seconds in the past. For example:

> BACKUP...AS OF SYSTEM TIME '-10s';

This improves performance by decreasing the likelihood that the BACKUP will be retried because it contends with other statements/transactions. However, because AS OF SYSTEM TIME returns historical data, your reads might be stale.

Automating backups

We recommend automating daily backups of your cluster.

To automate backups, you must have a client send the BACKUP statement to the cluster.

Once the backup is complete, your client will receive a BACKUP response.

Viewing and controlling backups jobs

After CockroachDB successfully initiates a backup, it registers the backup as a job, which you can view with SHOW JOBS.

After the backup has been initiated, you can control it with PAUSE JOB, RESUME JOB, and CANCEL JOB.

Synopsis

Required privileges

Only members of the admin role can run BACKUP. By default, the root user belongs to the admin role.

Parameters

Backup file URLs

We will use the URL provided to construct a secure API call to the service you specify. The path to each backup must be unique, and the URL for your backup's destination/locations must use the following format:

[scheme]://[host]/[path]?[parameters]

• ¹ If the AUTH parameter is implicit, all GCS connections use Google's default authentication strategy. If the AUTH parameter is default, the cloudstorage.gs.default.key cluster setting must be set to the contents of a service account file which will be used during authentication. If the AUTH parameter is not specified, the cloudstorage.gs.default.key setting will be used if it is non-empty, otherwise the implicit behavior is used.

• ² You can create your own HTTP server with Caddy or nginx. A custom root CA can be appended to the system's default CAs by setting the cloudstorage.http.custom_ca cluster setting, which will be used when verifying certificates from HTTPS URLs.

• ³ The file system backup location on the NFS drive is relative to the path specified by the --external-io-dir flag set while starting the node. If the flag is set to disabled, then imports from local directories and NFS drives are disabled.

• ⁴ A custom root CA can be appended to the system's default CAs by setting the cloudstorage.http.custom_ca cluster setting, which will be used when verifying certificates from an S3-compatible service.

• ⁵ The AWS_REGION parameter is optional since it is not a required parameter for most S3-compatible services. Specify the parameter only if your S3-compatible service requires it.

Example file URLs

Examples

Per our guidance in the Performance section, we recommend starting backups from a time at least 10 seconds in the past using AS OF SYSTEM TIME.

Backup a single table or view

> BACKUP bank.customers \
TO 'gs://acme-co-backup/database-bank-2017-03-27-weekly' \
AS OF SYSTEM TIME '-10s';

Backup multiple tables

> BACKUP bank.customers, bank.accounts \
TO 'gs://acme-co-backup/database-bank-2017-03-27-weekly' \
AS OF SYSTEM TIME '-10s';

Backup an entire database

> BACKUP DATABASE bank \
TO 'gs://acme-co-backup/database-bank-2017-03-27-weekly' \
AS OF SYSTEM TIME '-10s';

Backup with revision history

> BACKUP DATABASE bank \
TO 'gs://acme-co-backup/database-bank-2017-03-27-weekly' \
AS OF SYSTEM TIME '-10s' WITH revision_history;

Create incremental backups

Incremental backups must be based off of full backups you've already created.

> BACKUP DATABASE bank \
TO 'gs://acme-co-backup/db/bank/2017-03-29-nightly' \
AS OF SYSTEM TIME '-10s' \
INCREMENTAL FROM 'gs://acme-co-backup/database-bank-2017-03-27-weekly', 'gs://acme-co-backup/database-bank-2017-03-28-nightly';

Create incremental backups with revision history

> BACKUP DATABASE bank \
TO 'gs://acme-co-backup/database-bank-2017-03-29-nightly' \
AS OF SYSTEM TIME '-10s' \
INCREMENTAL FROM 'gs://acme-co-backup/database-bank-2017-03-27-weekly', 'gs://acme-co-backup/database-bank-2017-03-28-nightly' WITH revision_history;

See also

• RESTORE
• Configure Replication Zones

Was this helpful?