Creating datacenter backups is an important step in production deployments. Backups provide a mechanism for the Consul server to recover from an outage (network loss, operator error, or a corrupted data directory). All servers write to the `-data-dir` before commit on write requests. The same directory is used on client agents to persist local state too, but this is not critical and can be rebuilt when recreating an agent. Local client state is not backed up in this guide and doesn't need to be in general, only the server's Raft store state.
Consul provides the [snapshot](https://consul.io/docs/commands/snapshot.html) command which can be run using the CLI or the API. The `snapshot` command saves a point-in-time snapshot of the state of the Consul servers which includes, but is not limited to:
With [Consul Enterprise](/docs/commands/snapshot/agent), the `snapshot agent` command runs periodically and writes to local or remote storage (such as Amazon S3).
By default, all snapshots are taken using `consistent` mode where requests are forwarded to the leader which verifies that it is still in power before taking the snapshot. Snapshots will not be saved if the datacenter is degraded or if no leader is available. To reduce the burden on the leader, it is possible to [run the snapshot](/docs/commands/snapshot/save) on any non-leader server using `stale` consistency mode.
This spreads the load across nodes at the possible expense of losing full consistency guarantees. Typically this means that a very small number of recent writes may not be included. The omitted writes are typically limited to data written in the last `100ms` or less from the recovery point. This is usually suitable for disaster recovery. However, the system can’t guarantee how stale this may be if executed against a partitioned server.
The `snapshot save` command for backing up the datacenter state has many configuration options. In a production environment, you will want to configure ACL tokens and client certificates for security. The configuration options also allow you to specify the datacenter and server to collect the backup data from. Below are several examples.
To understand each field review the inspect [documentation](https://www.consul.io/docs/commands/snapshot/inspect.html). Notably, the `Version` field does not correspond to the version of the data. Rather it is the snapshot format version.
For production use, the `snapshot save` command or [API](https://www.consul.io/api/snapshot.html) should be scripted and run frequently. In addition to frequently backing up the datacenter state, there are several use cases when you would also want to manually execute `snapshot save`. First, you should always backup the datacenter before upgrading. If the upgrade does not go according to plan it is often not possible to downgrade due to changes in the state store format. Restoring from a backup is the only option so taking one before the upgrade will ensure you have the latest data. Second, if the datacenter loses quorum it may be beneficial to save the state before the servers become divergent. Finally, you can manually snapshot a datacenter and use that to bootstrap a new datacenter with the same state.
Operationally, the backup process does not need to be executed on every server. Additionally, you can use the configuration options to save the backups to a mounted filesystem. The mounted filesystem can even be cloud storage, such as Amazon S3. The enterprise command `snapshot agent` automates this process.
Running the `restore` process should be straightforward. However, there are a couple of actions you can take to ensure the process goes smoothly. First, make sure the datacenter you are restoring is stable and has a leader. You can see this using `consul operator raft list-peers` and checking server logs and telemetry for signs of leader elections or network issues.
In this guide, we learned about the `snapshot save` and `snapshot restore` commands. If you are testing the backup and restore process, you can add an extra dummy value to Consul KV. Another indicator that the backup was saved correctly is the size of the backup artifact.
