BACKUP

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.

Note:

The syntax BACKUP ... INTO adds a backup to a collection within the backup destination. The path to the backup is created using a date-based naming scheme. Versions of CockroachDB prior to v21.1 used the syntax BACKUP ... TO to backup directly to a specific operator-chosen destination, rather than picking a date-based path. The BACKUP ... TO syntax will be deprecated in future releases. For more information on this soon-to-be deprecated syntax, see the docs for v20.2 or earlier.

Note:

Core users can only take full backups. To use the other backup features, you need an Enterprise license. You can also use CockroachDB Dedicated, which runs full backups daily and incremental backups hourly.

You can backup a full cluster, which includes:

You can also backup:

  • An individual database, which includes all of its tables and views
  • An individual table, which includes its indexes and views

    BACKUP only backs up entire tables; it does not support backing up subsets of a table.

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.

Tip:

To view the contents of an Enterprise backup created with the BACKUP statement, use SHOW BACKUP.

Note:

BACKUP is a blocking statement. To run a backup job asynchronously, use the DETACHED option. See the options below.

Required privileges

  • Full cluster backups can only be run by members of the admin role. By default, the root user belongs to the admin role.
  • For all other backups, the user must have read access on all objects being backed up. Database backups require CONNECT privileges, and table backups require SELECT privileges. Backups of user-defined schemas, or backups containing user-defined types, require USAGE privileges.
  • BACKUP requires full read and write (including delete and overwrite) permissions to its target destination.

Destination privileges

The destination file URL does not require the ADMIN role in the following scenarios:

  • S3 and GS using SPECIFIED (and not IMPLICIT) credentials. Azure is always SPECIFIED by default.
  • Userfile

The destination file URL does require the ADMIN role in the following scenarios:

We recommend using cloud storage for bulk operations.

Synopsis

BACKUP TABLE table_pattern , DATABASE database_name , INTO subdirectory LATEST IN destination ( partitioned_backup_location , ) AS OF SYSTEM TIME timestamp WITH backup_options , OPTIONS ( backup_options , )

Parameters

Parameter Description
targets Back up the listed targets.
subdirectory The name of the specific subdirectory (e.g., 2021/03/23-213101.37) where you want to add an incremental backup. To view available subdirectories, use SHOW BACKUPS IN destination. If the subdirectory is not provided, a full backup will be created in the collection using a date-based naming scheme (i.e., <year>/<month>/<day>-<timestamp>).

Warning: If you use an arbitrary STRING as the subdirectory, a new full backup will be created, but it will never be shown in SHOW BACKUPS IN. We do not recommend using arbitrary strings as subdirectory names.
LATEST Append an incremental backup to the latest completed full backup's subdirectory.
destination The URL where you want to store the backup.

For information about this URL structure, see Backup File URLs.
timestamp Back up data as it existed as of timestamp. The timestamp must be more recent than your cluster's last garbage collection (which defaults to occur every 25 hours, but is configurable per table).
backup_options Control the backup behavior with a comma-separated list of these options.

Targets

Target Description
N/A Backup the cluster. For an example of a full cluster backup, see Backup a cluster.
DATABASE {database_name} [, ...] The name of the database(s) you want to backup (i.e., create backups of all tables and views in the database). For an example of backing up a database, see Backup a database.
TABLE {table_name} [, ...] The name of the table(s) or view(s) you want to backup. For an example of backing up a table or view, see Backup a table or view.

Options

Option Value Description
revision_history N/A Create a backup with full revision history, which records every change made to the cluster within the garbage collection period leading up to and including the given timestamp.
encryption_passphrase STRING The passphrase used to encrypt the files (BACKUP manifest and data files) that the BACKUP statement generates. This same passphrase is needed to decrypt the file when it is used to restore and to list the contents of the backup when using SHOW BACKUP. There is no practical limit on the length of the passphrase.
DETACHED N/A When a backup runs in DETACHED mode, it will execute asynchronously and the job ID will be returned immediately without waiting for the job to finish. Note that with DETACHED specified, further job information and the job completion status will not be returned. For more on the differences between the returned job data, see the example below. To check on the job status, use the SHOW JOBS statement.

To run a backup within a transaction, use the DETACHED option.
kms STRING The key management service (KMS) URI (or a comma-separated list of URIs) used to encrypt the files (BACKUP manifest and data files) that the BACKUP statement generates. This same KMS URI is needed to decrypt the file when it is used to restore and to list the contents of the backup when using SHOW BACKUP.

Currently, only AWS KMS is supported.

Backup file URLs

CockroachDB uses the URL provided to construct a secure API call to the service you specify. The URL structure depends on the type of file storage you are using. For more information, see the following:

Functional details

Object dependencies

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

Object Depends On
Table with foreign key constraints The table it REFERENCES; however, this dependency can be removed during the restore.
Table with a sequence The sequence it uses; however, this dependency can be removed during the restore.
Views The tables used in the view's SELECT statement.

Users and privileges

The system.users table stores your users and their passwords. To restore your users and privilege grants, do a cluster backup and restore the cluster to a fresh cluster with no user data. You can also backup the system.users table, and then use this procedure.

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.

BACKUP, like any read, cannot export a range if the range contains an unresolved intent. While you typically will want bulk, background jobs like BACKUP to have as little impact on your foreground traffic as possible, it's more important for backups to actually complete (which maintains your recovery point objective (RPO)). Unlike a normal read transaction that will block until any uncommitted writes it encounters are resolved, BACKUP will block only for a configurable duration before invoking priority to ensure it can complete on-time.

We 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. Taking backups with AS OF SYSTEM TIME '-10s' is a good best practice to reduce the number of still-running transactions you may encounter, since the backup will take priority and will force still-running transactions to restart after the backup is finished.

BACKUP will initially ask individual ranges to backup but to skip if they encounter an intent. Any range that is skipped is placed at the end of the queue. When BACKUP has completed its initial pass and is revisiting ranges, it will ask any range that did not resolve within the given time limit (default 1 minute) to attempt to resolve any intents that it encounters and to not skip. Additionally, the backup's transaction priority is then set to high, which causes other transactions to abort until the intents are resolved and the backup is finished.

Note:

To set a target for the amount of backup data written to each backup file, use the bulkio.backup.file_size cluster setting.

See the SET CLUSTER SETTING page for more details on using cluster settings.

Viewing and controlling backups jobs

After CockroachDB successfully initiates a backup, it registers the backup as a job, and you can do the following:

Action SQL Statement
View the backup status SHOW JOBS
Pause the backup PAUSE JOB
Resume the backup RESUME JOB
Cancel the backup CANCEL JOB

You can also visit the Jobs page of the DB Console to view job details. The BACKUP statement will return when the backup is finished or if it encounters an error.

Note:

The presence of a BACKUP-CHECKPOINT file in the backup destination usually means the backup is not complete. This file is created when a backup is initiated, and is replaced with a BACKUP file once the backup is finished.

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. Each example below follows this guidance.

The examples below provide connection strings to Amazon S3, Google Cloud Storage, and Azure Storage. For guidance on connecting to other storage options or using other authentication parameters, read Use Cloud Storage for Bulk Operations.

Note:

The examples in this section use the default AUTH=specified parameter. For more detail on how to use implicit authentication with Amazon S3 buckets, read Use Cloud Storage for Bulk Operations — Authentication.

Backup a cluster

To take a full backup of a cluster:

icon/buttons/copy
> BACKUP INTO \
's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a database

To take a full backup of a single database:

icon/buttons/copy
> BACKUP DATABASE bank \
INTO 's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple databases:

icon/buttons/copy
> BACKUP DATABASE bank, employees \
INTO 's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a table or view

To take a full backup of a single table or view:

icon/buttons/copy
> BACKUP bank.customers \
INTO 's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple tables:

icon/buttons/copy
> BACKUP bank.customers, bank.accounts \
INTO 's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';
Warning:

CockroachDB does not currently support BACKUPs of individual tables that are part of multi-region databases. For example, you cannot backup a GLOBAL or REGIONAL table individually.

To work around this limitation, you must back up the database or the entire cluster.

Tracking GitHub Issue

Backup all tables in a schema

New in v21.2: To back up all tables in a specified schema, use a wildcard with the schema name:

icon/buttons/copy
> BACKUP test_schema.*
INTO 's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

Alternatively, use a fully qualified name: database.schema.*.

With this syntax, schemas will be resolved before databases. test_object.* will resolve to a schema of test_object within the set current database before matching to a database of test_object.

If a database and schema have the same name, such as bank.bank, running BACKUP bank.* will result in the schema resolving first. All the tables within that schema will be backed up. However, if this were to be run from a different database that does not have a bank schema, all tables in the bank database will be backed up.

See Name Resolution for more details on how naming hierarchy and name resolution work in CockroachDB.

Create incremental backups

If you backup to a destination already containing a full backup, an incremental backup will be appended to the full backup's path with a date-based name (e.g., 20210324):

icon/buttons/copy
> BACKUP INTO LATEST IN \
's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s';

Run a backup asynchronously

Use the DETACHED option to execute the backup job asynchronously:

icon/buttons/copy
> BACKUP INTO \
's3://{BUCKET NAME}/{PATH}?AWS_ACCESS_KEY_ID={KEY ID}&AWS_SECRET_ACCESS_KEY={SECRET ACCESS KEY}' \
AS OF SYSTEM TIME '-10s'
WITH DETACHED;

The job ID is returned immediately without waiting for the job to finish:

        job_id
----------------------
  592786066399264769
(1 row)

Without the DETACHED option, BACKUP will block the SQL connection until the job completes. Once finished, the job status and more detailed job data is returned:

job_id             |  status   | fraction_completed | rows | index_entries | bytes
---------------------+-----------+--------------------+------+---------------+--------
652471804772712449 | succeeded |                  1 |   50 |             0 |  4911
(1 row)

Backup a cluster

To take a full backup of a cluster:

icon/buttons/copy
> BACKUP INTO \
'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a database

To take a full backup of a single database:

icon/buttons/copy
> BACKUP DATABASE bank \
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple databases:

icon/buttons/copy
> BACKUP DATABASE bank, employees \
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a table or view

To take a full backup of a single table or view:

icon/buttons/copy
> BACKUP bank.customers \
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple tables:

icon/buttons/copy
> BACKUP bank.customers, bank.accounts \
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';
Warning:

CockroachDB does not currently support BACKUPs of individual tables that are part of multi-region databases. For example, you cannot backup a GLOBAL or REGIONAL table individually.

To work around this limitation, you must back up the database or the entire cluster.

Tracking GitHub Issue

Backup all tables in a schema

New in v21.2: To back up all tables in a specified schema, use a wildcard with the schema name:

icon/buttons/copy
> BACKUP test_schema.*
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Alternatively, use a fully qualified name: database.schema.*.

With this syntax, schemas will be resolved before databases. test_object.* will resolve to a schema of test_object within the set current database before matching to a database of test_object.

If a database and schema have the same name, such as bank.bank, running BACKUP bank.* will result in the schema resolving first. All the tables within that schema will be backed up. However, if this were to be run from a different database that does not have a bank schema, all tables in the bank database will be backed up.

See Name Resolution for more details on how naming hierarchy and name resolution work in CockroachDB.

Create incremental backups

If you backup to a destination already containing a full backup, an incremental backup will be appended to the full backup's path with a date-based name (e.g., 20210324):

icon/buttons/copy
> BACKUP INTO LATEST IN \
'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Run a backup asynchronously

Use the DETACHED option to execute the backup job asynchronously:

icon/buttons/copy
> BACKUP INTO \
'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s'
WITH DETACHED;

The job ID is returned immediately without waiting for the job to finish:

        job_id
----------------------
  592786066399264769
(1 row)

Without the DETACHED option, BACKUP will block the SQL connection until the job completes. Once finished, the job status and more detailed job data is returned:

job_id             |  status   | fraction_completed | rows | index_entries | bytes
---------------------+-----------+--------------------+------+---------------+--------
652471804772712449 | succeeded |                  1 |   50 |             0 |  4911
(1 row)
Note:

The examples in this section use the AUTH=specified parameter, which will be the default behavior in v21.2 and beyond for connecting to Google Cloud Storage. For more detail on how to pass your Google Cloud Storage credentials with this parameter, or, how to use implicit authentication, read Use Cloud Storage for Bulk Operations — Authentication.

Backup a cluster

To take a full backup of a cluster:

icon/buttons/copy
> BACKUP INTO \
'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a database

To take a full backup of a single database:

icon/buttons/copy
> BACKUP DATABASE bank \
INTO 'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple databases:

icon/buttons/copy
> BACKUP DATABASE bank, employees \
INTO 'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Backup a table or view

To take a full backup of a single table or view:

icon/buttons/copy
> BACKUP bank.customers \
INTO 'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

To take a full backup of multiple tables:

icon/buttons/copy
> BACKUP bank.customers, bank.accounts \
INTO 'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';
Warning:

CockroachDB does not currently support BACKUPs of individual tables that are part of multi-region databases. For example, you cannot backup a GLOBAL or REGIONAL table individually.

To work around this limitation, you must back up the database or the entire cluster.

Tracking GitHub Issue

Backup all tables in a schema

New in v21.2: To back up all tables in a specified schema, use a wildcard with the schema name:

icon/buttons/copy
> BACKUP test_schema.*
INTO 'azure://{CONTAINER NAME}/{PATH}?AZURE_ACCOUNT_NAME={ACCOUNT NAME}&AZURE_ACCOUNT_KEY={URL-ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Alternatively, use a fully qualified name: database.schema.*.

With this syntax, schemas will be resolved before databases. test_object.* will resolve to a schema of test_object within the set current database before matching to a database of test_object.

If a database and schema have the same name, such as bank.bank, running BACKUP bank.* will result in the schema resolving first. All the tables within that schema will be backed up. However, if this were to be run from a different database that does not have a bank schema, all tables in the bank database will be backed up.

See Name Resolution for more details on how naming hierarchy and name resolution work in CockroachDB.

Create incremental backups

If you backup to a destination already containing a full backup, an incremental backup will be appended to the full backup's path with a date-based name (e.g., 20210324):

icon/buttons/copy
> BACKUP INTO LATEST IN \
'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s';

Run a backup asynchronously

Use the DETACHED option to execute the backup job asynchronously:

icon/buttons/copy
> BACKUP INTO \
'gs://{BUCKET NAME}/{PATH}?AUTH=specified&CREDENTIALS={ENCODED KEY}' \
AS OF SYSTEM TIME '-10s'
WITH DETACHED;

The job ID is returned immediately without waiting for the job to finish:

        job_id
----------------------
  592786066399264769
(1 row)

Without the DETACHED option, BACKUP will block the SQL connection until the job completes. Once finished, the job status and more detailed job data is returned:

job_id             |  status   | fraction_completed | rows | index_entries | bytes
---------------------+-----------+--------------------+------+---------------+--------
652471804772712449 | succeeded |                  1 |   50 |             0 |  4911
(1 row)

Advanced examples

For examples of advanced BACKUP and RESTORE use cases, see:

Known limitations

BACKUP of multi-region tables

CockroachDB does not currently support BACKUPs of individual tables that are part of multi-region databases. For example, you cannot backup a GLOBAL or REGIONAL table individually.

To work around this limitation, you must back up the database or the entire cluster.

Tracking GitHub Issue

See also

YesYes NoNo