Payments are experiencing issues due to temporary restrictions in Russia. If your payment does not go through, please submit a support request.Our support team is available 24/7 — we are always here to help with hosting and server issues.We are now accepting requests for dedicated server rental and colocation services in our data center.Reminder: we recommend enabling backups for additional data protection.A new VPS/VDS lineup with NVMe storage and improved performance is now available.Maintenance work on some servers has been completed. All services are operating normally.
Article5 min readViews0

Encrypted backup is in place, but there is no access to the data: check your secrets

We separate storage access rights from decryption capabilities using restic as an example, look for circular dependencies, and verify emergency access.

Comments 0

Mikhail checks an external drive next to a sealed emergency envelope: data and decryption access
In this article

Backup files are accessible, their size looks normal, but the program cannot decrypt the data. Such a failure does not require losing the storage itself. It is sufficient to lose the last valid way to open the encrypted repository. Therefore, access to the storage location and access to the copy contents must be checked separately.

Let us examine this distinction using restic, based on version 0.19.1 documentation read on September 27, 2026. This is a readiness verification model, without commands to create a repository, change passwords, or delete keys. The described recovery on a test stand was not performed here. The team must obtain its own result on an authorized isolated copy.

Two different permissions

Storage credentials determine the ability to retrieve repository files. The restic repository password is required to access protected data. In the documentation for remote storage, these mechanisms are described separately: for example, access rights to an object storage do not replace the password requested by the backup program.

Scenario: after a server loss, the administrator restored access to the storage account and downloaded files. The only repository password remained in the configuration on the lost server. A successful download in this case confirms only the first stage. It does not prove that the data can be read.

The reverse situation is also possible: the password is known, but the storage account, second factor, or read permission is unavailable. Here the problem lies before decryption. These cases require different actions, so the generic message "no access to backup" should be clarified before contacting support.

Recovery requires separate access to storage files and the ability to decrypt the repository
Storage credentials and the repository password solve different problems. Having one does not replace the other.

What it means to lose the last decryption method

restic allows multiple access keys or passwords for a single repository. Therefore, losing one entry does not necessarily mean losing data if another valid and verified access method remains. However, relying on such a backup is possible only after confirming that it belongs to the correct repository and actually works.

The restic documentation warns that data cannot be recovered if the required password is lost. In a practical plan, this means at least one valid decryption method must be preserved independently of the failed machine. Restoring the storage provider account should not be considered a restoration of the encryption secret.

This conclusion does not require weakening copy protection or making the password publicly available. A separate authorized procedure for obtaining the secret with a designated responsible person is needed. The general instruction specifies the location and access procedure, but the secret itself is not included in it.

Find circular dependencies

Imagine that the backup password is stored in a secrets manager, while the only backup of the secrets manager itself is encrypted in the same inaccessible way. On the diagram, both elements exist, but there is nothing to start the recovery from. A similar dependency arises when login confirmation is possible only via a lost phone belonging to a single employee.

The practical question is: can the designated specialist start the procedure from a clean authorized workstation if the original VPS and its local settings are no longer available? The answer must include access to the storage, a decryption tool, and a clear procedure for obtaining them. There is no need to simulate a real loss of access: the scenario is analyzed and tested in a specially prepared environment.

For each link, record the owner, the backup responsible person, and dependencies on other systems. Special attention is required for locations that are accessible only after the store being restored is launched. For example, an internal instruction existing only in its administrative part will not help until the site itself is restored.

Circular dependency: a password stored in a secrets manager, but its copy requires the same password; an independent authorized path is needed.
Dependency diagram. Emergency access must be authorized and verifiable without weakening copy protection.

What to check when changing access

After changing a password, key, or responsible person, it is important to verify not only the new automatic copy job. It may use the fresh secret on a still-running server, while the emergency procedure might reference an outdated record. A consistency check of both paths is required: the standard copy creation process and independent restoration.

The acceptance test plan includes selecting the appropriate repository and point, obtaining authorized access from the prepared environment, reading the data, and verifying the restored test dataset. Opening a file list is insufficient to confirm that the entire store will be restored. Data composition, application versions, and integration functionality must be checked separately.

Working passwords should not be added to screenshots, reports, command history, or correspondence. Confirmation is sufficient by specifying which authorized access method was verified, by whom, when, and with what result. If a secret cannot be obtained, this is a specific incomplete part of the test, not a reason to mark the copy as valid based solely on archive size.

For an encrypted backup, both the data and the decryption path must survive an incident. When these two dependencies are verified separately, the message "files are in place" gains useful context: it becomes known who and by which verified method can open them.

Discussion 0

Share your experience and ask questions. Comments without links appear after editorial review.

No comments yet. Start the discussion.