Guidelines for device backup and restore

Caution

Users with CLI access can access the Linux shell access with the expert command, which can present a security risk. For system security reasons, we strongly recommend:

  • Only use the Linux shell under TAC supervision or when explicitly instructed by Firewall and Cloud-Delivered Firewall Management Center user documentation.

  • Restrict the list of users with Linux shell access.

  • Do not add users directly in the Linux shell; only use the procedures in this chapter.

Backup and restore is for device Recovery/Return Material Authorization

Backup and restore are primarily intended for Return Material Authorization (RMA) scenarios. Do not execute the restore process on a faulty or failed physical device. Execute the process only on a replaced hardware.

If the backup resides only on the faulty device, it will be erased when you reimage the device. Similarly, if the device fails to boot up, you may not be able to recover the backup. Hence, ensure that you always secure a copy of the backup that can be accessed for restoration.

Device backup and restore does not equate to import/export of device configuration

A backup file contains information that uniquely identifies a device that cannot be shared. Do not use the backup and restore process to copy configurations between devices, or as a way to save configurations while testing new ones. Instead, use the import/export feature.

For example, Firewall Threat Defense device backups include the device's management IP address and all information needed for the device to connect to its managing Cloud-Delivered Firewall Management Center. Do not restore the Firewall Threat Defense backup to a device being managed by a different manager; the restored device attempts to connect only to the manager specified in the backup.

Restore is individual and local

You restore threat defense devices individually and locally. This means:

  • You cannot use Cloud-Delivered Firewall Management Center to restore a device. You must use the Firewall Threat Defense CLI to restore. However, for the ISA 3000 zero-touch restore, you would use an SD card and the reset button.

  • You cannot use a Cloud-Delivered Firewall Management Center user account to log into and restore one of its managed devices. Cloud-Delivered Firewall Management Center and Firewall Threat Defense devices maintain their own user accounts.

SSH host key verification for remote SCP storage

The Firewall Threat Defense verifies the SSH host key of a remote SCP storage server before starting file transfers. Previously, SCP operations could connect without requiring the remote server key in the local /root/.ssh/known_hostsfile.

Administrators must add and verify the remote storage server’s SSH host key before using remotely stored backup for Firewall Threat Defense restore.

Before adding a key, obtain or confirm the expected SSH host-key fingerprint with the remote storage server administrator.

If the host key is missing or does not match the key presented by the remote server, the transfer is blocked. Typical error messages include Host key verification failed and Error in copying backup file from remote location.

On the replacement Firewall Threat Defense CLI console, connect to the SSH server once and accept the new key:

ssh -o StrictHostKeyChecking=accept-new -o ConnectTimeout=10 @ exit

Verify the stored key:

ssh-keygen -F -l

Use the same IP address or hostname in the known_hosts entry as used in the backup, restore, or migration configuration. If the remote server’s SSH host key changes legitimately, remove the old entry, confirm the new fingerprint with the server administrator, and add the new key before retrying the operation.

Backup and restore in Firewall Threat Defense high availability deployments

In a Firewall Threat Defense HA deployment, you must:

  • Back up the device pair from the Cloud-Delivered Firewall Management Center, but restore individually and locally from the Firewall Threat Defense CLI.

    The backup process produces unique backup files for Firewall Threat Defense HA devices. Do not restore one HA peer with the backup file from the other. A backup file contains information that uniquely identifies an appliance, and cannot be shared.

    A Firewall Threat Defense HA device's role is noted in its backup file name. When you restore, make sure you choose the appropriate backup file: primary or secondary.

  • Do not suspend or break HA before you restore.

    Maintaining the HA configuration ensures that replacement devices can easily reconnect after restore. Note that you will have to resume HA synchronization to make this happen.

  • Do not run the restore CLI command on both peers at the same time.

    Assuming you have successful backups, you can replace either or both peers in an HA pair. Any physical replacement tasks that you can perform simultaneously: unracking, reracking, and so on. However, do not run the restore command on the second device until the restore process completes for the first device, including the reboot.

Backup and Restore in Firewall Threat Defense Clustering Deployments

In the Firewall Threat Defense clustering deployment, you must:

  • Back up the entire cluster from the Cloud-delivered Firewall Management Center, but restore nodes individually and locally from the threat defense CLI.

    The backup process produces unique backup files for each cluster node. Do not restore one node with the backup file from another. A backup file contains information that uniquely identifies a device, and cannot be shared.

    The node's role is noted in its backup file name. When you restore, make sure you choose the appropriate backup file: control or data.

    If the control node fails a pre-backup check, backups of its data nodes are also prevented from starting. If the control node passes these checks but its backup fails later, data node backups continue. A data node backup failure does not automatically stop the other node backups. Check the backup result for each node.

  • All the nodes that are part of the cluster must be registered in the management center, for the backup to be successful.

  • Do not suspend or break clustering before you restore. Maintaining the cluster configuration ensures replacement devices can easily reconnect after restore.

  • Do not run the restore CLI command on multiple nodes at the same time. We recommend that you restore the control node first and wait until it rejoins the cluster before you restore any data nodes.

    Assuming you have successful backups, you can replace multiple nodes in the cluster. Any physical replacement tasks you can perform simultaneously: unracking, reracking, and so on. However, do not run the restore command on an additional node until the restore process completes for the previous node, including the reboot.