Restore data

Use the command to restore backup data and metadata from InfluxDB OSS.

InfluxDB moves existing data and metadata to a temporary location. If the restore fails, InfluxDB preserves temporary data for recovery, otherwise this data is deleted. See .

Cannot restore to existing buckets

The influx restore command cannot restore data to existing buckets. Use the --new-bucket flag to create a new bucket to restore data to. To restore data and retain bucket names, and then begin the restore process.

Use the influx restore command and specify the path to the backup directory.

For more information about restore options and flags, see the influx restore documentation.

To restore data from a specific backup bucket, provide the following:

  • backup directory path
  • bucket name or ID
  1. influx restore \
  2.   /backups/2020-01-20_12-00/ \
  3.   --bucket example-bucket
  5. # OR
  7. influx restore \
  8.   /backups/2020-01-20_12-00/ \
  9.   --bucket-id 000000000000

If a bucket with the same name as the backed up bucket already exists in InfluxDB, use the --new-bucket flag to create a new bucket with a different name and restore data into it.

To restore and replace all time series data and InfluxDB key-value data such as tokens, users, dashboards, etc., include the following:

  • --full flag
  • backup directory path
  1.   /backups/2020-01-20_12-00/ \
  2.   --full

Restore to a new InfluxDB server

If using a backup to populate a new InfluxDB server:

  1. Retrieve the from your source InfluxDB instance.

  3. Restore the backup to the new server.

    1. influx restore \
    2.   /backups/2020-01-20_12-00/ \
    3.   --full

If you do not provide the admin token from your source InfluxDB instance as the admin token in your new instance, the restore process and all subsequent attempts to authenticate with the new server will fail.

  1. The first restore API call uses the auto-generated token to authenticate with the new server and overwrites the entire key-value store in the new server, including the auto-generated token.
  2. The second restore API call attempts to upload time series data, but uses the auto-generated token to authenticate with new server. That token is overwritten in first restore API call and the process fails to authenticate.

Recover from a failed restore

If the restoration process fails, InfluxDB preserves existing data in a tmp directory in the (default is ~/.influxdbv2/engine).

To recover from a failed restore:

  1. Copy the temporary files back into the engine directory.
  2. Remove the .tmp extensions from each of the copied files.
  3. Restart the server.