Backup and restore a database

You can use Database services to backup a database and store the backup artifact in the Object Storage service. Later on, if the original database is damaged, you can use the backup artifact to restore the database. The restore process creates a new database instance.

The backup data is stored in OpenStack Swift, the user is able to customize which container to store the data. The following ways are described in the order of precedence from greatest to least:

1. The container name can be specified when creating backups, this could override either the backup strategy setting or the default setting in Trove configuration.

2. Users could create backup strategy either for the project scope or for a particular instance.

3. If not configured by the end user, will use the default value in Trove configuration.

Caution

If the objects in the backup container are manually deleted, the database can’t be properly restored.

This example shows you how to create backup strategy, create backup and restore instance from the backup.

1. Before creating backup

   1. Make sure you have created an instance, e.g. in this example, we use the following instance:

      $ openstack database instance list
      +--------------------------------------+--------+-----------+-------------------+--------+-----------+------+
      |                  id                  |  name  | datastore | datastore_version | status | flavor_id | size |
      +--------------------------------------+--------+-----------+-------------------+--------+-----------+------+
      | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 | guest1 |   mysql   |     mysql-5.5     | ACTIVE |     10    |  2   |
      +--------------------------------------+--------+-----------+-------------------+--------+-----------+------+
      

   2. Optionally, create a backup strategy for the instance. You can also specify a different swift container name (--swift-container) when creating the backup.

      $ openstack database backup strategy create --instance-id 97b4b853-80f6-414f-ba6f-c6f455a79ae6 --swift-container my-trove-backups
      +-----------------+--------------------------------------+
      | Field           | Value                                |
      +-----------------+--------------------------------------+
      | backend         | swift                                |
      | instance_id     | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 |
      | project_id      | 922b47766bcb448f83a760358337f2b4     |
      | swift_container | my-trove-backups                     |
      +-----------------+--------------------------------------+
      

2. Backup the database instance

   Back up the database instance by using the openstack database backup create command. In this example, the backup is called backup1.

   $ openstack database backup create 97b4b853-80f6-414f-ba6f-c6f455a79ae6 backup1
   +-------------+--------------------------------------+
   |   Property  |                Value                 |
   +-------------+--------------------------------------+
   |   created   |         2014-03-18T17:09:07          |
   | description |                 None                 |
   |      id     | 8af30763-61fd-4aab-8fe8-57d528911138 |
   | instance_id | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 |
   | locationRef |                 None                 |
   |     name    |               backup1                |
   |  parent_id  |                 None                 |
   |     size    |                 None                 |
   |    status   |                 NEW                  |
   |   updated   |         2014-03-18T17:09:07          |
   +-------------+--------------------------------------+
   

   Later on, use either openstack database backup list command or openstack database backup show command to check the backup status:

   $ openstack database backup list
   +--------------------------------------+--------------------------------------+---------+-----------+-----------+---------------------+
   |                  id                  |             instance_id              |   name  |   status  | parent_id |       updated       |
   +--------------------------------------+--------------------------------------+---------+-----------+-----------+---------------------+
   | 8af30763-61fd-4aab-8fe8-57d528911138 | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 | backup1 | COMPLETED |    None   | 2014-03-18T17:09:11 |
   +--------------------------------------+--------------------------------------+---------+-----------+-----------+---------------------+
   $ openstack database backup show 8af30763-61fd-4aab-8fe8-57d528911138
   +-------------+----------------------------------------------------+
   |   Property  |                   Value                            |
   +-------------+----------------------------------------------------+
   |   created   |              2014-03-18T17:09:07                   |
   | description |                   None                             |
   |      id     |                 8af...138                          |
   | instance_id |                 97b...ae6                          |
   | locationRef | http://10.0.0.1:.../.../8af...138.xbstream.gz.enc  |
   |     name    |                 backup1                            |
   |  parent_id  |                  None                              |
   |     size    |                  0.17                              |
   |    status   |               COMPLETED                            |
   |   updated   |           2014-03-18T17:09:11                      |
   +-------------+----------------------------------------------------+
   

3. Check the backup data in Swift

   Check the container is created and the backup data is saved as objects inside the container.

   $ openstack container list
   +------------------+
   | Name             |
   +------------------+
   | my-trove-backups |
   +------------------+
   $ openstack object list my-trove-backups
   +--------------------------------------------------+
   | Name                                             |
   +--------------------------------------------------+
   | 8af30763-61fd-4aab-8fe8-57d528911138.xbstream.gz |
   +--------------------------------------------------+
   

4. Restore a database instance

   Now assume that the guest1 database instance is damaged and you need to restore it. In this example, you use the openstack database instance create command to create a new database instance called guest2.

   • Specify that the new guest2 instance has the same flavor (10) and the same root volume size (2) as the original guest1 instance.

   • Use the --backup argument to indicate that this new instance is based on the backup artifact identified by BACKUP_ID. In this example, replace BACKUP_ID with 8af30763-61fd-4aab-8fe8-57d528911138.

   $ openstack database instance create guest2 --flavor 10 --size 2 --nic net-id=$network_id --backup BACKUP_ID
   +-------------------+----------------------------------------------+
   |      Property     |                Value                         |
   +-------------------+----------------------------------------------+
   |      created      |         2014-03-18T17:12:03                  |
   |     datastore     | {u'version': u'mysql-5.5', u'type': u'mysql'}|
   |datastore_version  |                mysql-5.5                     |
   |       flavor      | {u'id': u'10', u'links': [{u'href': ...]}    |
   |         id        |  ac7a2b35-a9b4-4ff6-beac-a1bcee86d04b        |
   |        name       |                guest2                        |
   |       status      |                 BUILD                        |
   |      updated      |          2014-03-18T17:12:03                 |
   |       volume      |             {u'size': 2}                     |
   +-------------------+----------------------------------------------+
   

5. Verify backup

   Now check that the new guest2 instance has the same characteristics as the original guest1 instance.

   Start by getting the ID of the new guest2 instance.

   $ openstack database instance list
   
   +-----------+--------+-----------+-------------------+--------+-----------+------+
   |     id    |  name  | datastore | datastore_version | status | flavor_id | size |
   +-----------+--------+-----------+-------------------+--------+-----------+------+
   | 97b...ae6 | guest1 |   mysql   |     mysql-5.5     | ACTIVE |     10    |  2   |
   | ac7...04b | guest2 |   mysql   |     mysql-5.5     | ACTIVE |     10    |  2   |
   +-----------+--------+-----------+-------------------+--------+-----------+------+
   

   Use the openstack database instance show command to display information about the new guest2 instance. Pass in guest2’s INSTANCE_ID, which is ac7a2b35-a9b4-4ff6-beac-a1bcee86d04b.

   $ openstack database instance show INSTANCE_ID
   +-------------------+--------------------------------------+
   |      Property     |                Value                 |
   +-------------------+--------------------------------------+
   |      created      |         2014-03-18T17:12:03          |
   |     datastore     |                mysql                 |
   | datastore_version |              mysql-5.5               |
   |       flavor      |                  10                  |
   |         id        | ac7a2b35-a9b4-4ff6-beac-a1bcee86d04b |
   |         ip        |               10.0.0.3               |
   |        name       |                guest2                |
   |       status      |                ACTIVE                |
   |      updated      |         2014-03-18T17:12:06          |
   |       volume      |                  2                   |
   |    volume_used    |                 0.18                 |
   +-------------------+--------------------------------------+
   

   Note that the data store, flavor ID, and volume size have the same values as in the original guest1 instance.

   Use the openstack database db list command to check that the original databases (db1 and db2) are present on the restored instance.

   $ openstack database db list INSTANCE_ID
   +--------------------+
   |        name        |
   +--------------------+
   |        db1         |
   |        db2         |
   | performance_schema |
   |        test        |
   +--------------------+
   

   Use the openstack database user list command to check that the original user (user1) is present on the restored instance.

   $ openstack database user list INSTANCE_ID
   +--------+------+-----------+
   |  name  | host | databases |
   +--------+------+-----------+
   | user1  |  %   |  db1, db2 |
   +--------+------+-----------+
   

6. Notify users

   Tell the users who were accessing the now-disabled guest1 database instance that they can now access guest2. Provide them with guest2’s name, IP address, and any other information they might need. (You can get this information by using the openstack database instance show command.)

7. Clean up

   At this point, you might want to delete the disabled guest1 instance, by using the openstack database instance delete command.

   $ openstack database instance delete INSTANCE_ID
   

Create incremental backups

Incremental backups let you chain together a series of backups. You start with a regular backup. Then, when you want to create a subsequent incremental backup, you specify the parent backup.

Restoring a database instance from an incremental backup is the same as creating a database instance from a regular backup. the Database service handles the process of applying the chain of incremental backups.

Create an incremental backup based on a parent backup:

$ openstack database backup create INSTANCE_ID backup1.1  --parent BACKUP_ID
+-------------+--------------------------------------+
|   Property  |                Value                 |
+-------------+--------------------------------------+
|   created   |         2014-03-19T14:09:13          |
| description |                 None                 |
|      id     | 1d474981-a006-4f62-b25f-43d7b8a7097e |
| instance_id | 792a6a56-278f-4a01-9997-d997fa126370 |
| locationRef |                 None                 |
|     name    |              backup1.1               |
|  parent_id  | 6dc3a9b7-1f3e-4954-8582-3f2e4942cddd |
|     size    |                 None                 |
|    status   |                 NEW                  |
|   updated   |         2014-03-19T14:09:13          |
+-------------+--------------------------------------+

Troubleshooting

Failed to create incremental backup for PostgreSQL

One possible reason could be it has been a long time since the parent backup was created, and the parent backup WAL file is removed internally because of disk pressure, it could be confirmed by checking the instance detail, e.g.

$ openstack database instance show e7231e46-ca3b-4dce-bf67-739b3af0ef85 -c fault
+-------+----------------------------------------------------------------------+
| Field | Value                                                                |
+-------+----------------------------------------------------------------------+
| fault | Failed to create backup c76de467-6587-4e27-bb8d-7c3d3b136663, error: |
|       |     Cannot find parent backup WAL file.                              |
+-------+----------------------------------------------------------------------+

In this case, you have to create full backup instead.

To avoid this issue in the future, you can set up a cron job to create (incremental) backups regularly.