Instruction 06-configuration-backups-and-recovery: Reading

A useful configuration backup is more than a compressed copy of a directory. A recovery operator must know what was captured, which system or service it belongs to, when it was captured, whether the archive remained intact, and whether the restored files are usable. These requirements create several separate validation layers. An archive listing proves that the container can be read, but it does not prove that the files express valid settings. A cryptographic checksum detects byte changes, but a perfectly intact file can still contain an invalid port, a missing required key, or a value incompatible with the installed application version. Syntax validation confirms that a parser accepts the document. Semantic validation then checks rules meaningful to the service, such as an integer being within an allowed range.

Recovery should be staged. Extracting directly over an active configuration mixes backup extraction, validation, and activation into one difficult-to-reverse operation. Instead, restore into a candidate directory, verify the archive checksum, verify the manifest, inspect metadata, parse the files, and apply semantic checks. Only after those gates pass should the candidate be promoted. Keeping the old configuration under a distinct name provides a rapid rollback path. Renaming directories on the same filesystem also reduces the interval in which an incomplete tree could be visible, although application behavior and open file handles still determine whether a restart or reload is required.

Checksums establish integrity relative to a trusted checksum, not authenticity by themselves. If an attacker can replace both an archive and its checksum, the pair can still verify. Higher-assurance environments should protect manifests with signing, immutable storage, restricted backup credentials, or an independently controlled backup system. Configuration backups may also contain passwords, API tokens, private keys, internal addresses, or identity-provider details. They therefore require access restrictions, retention limits, secure transfer, and tested destruction procedures.

The final measure of backup quality is a successful restore test. A job reporting that it created an archive only proves that a command completed. Regular drills should recover a representative version, validate it against the target software version, and document the time and decisions required. This lab models that workflow entirely under one classroom path: establish a known configuration, create a manifest and archive, introduce legitimate-looking drift, recover to an isolated candidate, validate it, promote it, and preserve a rollback route.