What it does
A scheduler cuts full and incremental slices of the log into an encrypted chain in a local store, verifies each by reading it back, and prunes what a restorable chain no longer needs. restore-to-seq rebuilds a fresh data directory at a chosen position, and a drill restores the latest verified cover and compares it with an independent replay.
How it works
kynetra backup schedule runs against a local chain store directory with a durable state file. Each tick decides whether a new full slice is due, because none exists, the full cadence elapsed or the incremental chain is long, or whether the recovery-point window has elapsed since the last verified cover. If so it reads the relevant range of the log, seals it into a manifest and an AES-256-GCM encrypted payload, writes it to the store, and verifies it by reading it back, checksumming and decrypting it. A bundle advances the recorded recovery point when verification passes, not before, so a failed tick shows up at once as a missed target and the next tick retries. The defaults are a 300 second recovery point, a 900 second restore time, a full slice every 86,400 seconds, three fulls and 288 incrementals kept; all are flags. --once ticks a single time for cron or a timer, and a failed tick exits non-zero.
A chain bundle carries raw log records rather than reconstructed entities, so a full slice and its incrementals replay onto a fresh engine with the original sequence numbers. Incrementals are contiguous by construction, and a chain read back from disk is checked for gaps, reordered links, a tampered range and the wrong key. A full capture is refused if the source log has been trimmed, because such a bundle would verify perfectly and still be unrestorable.
kynetra backup restore-to-seq materialises the chain covering a target position into a new, empty directory. The target is inclusive, and a target below the oldest retained base or beyond coverage is refused. A randomised corpus of restore times checks that the restored state reaches exactly the requested commit with no later writes, in both entity profiles.
kynetra backup drill restores the latest verified cover into a scratch directory, compares the result with a second engine built by replaying the source log directly, measures the time against the restore-time target, and writes a report. kynetra backup rpo-status prints the scheduler's recorded recovery point from its state file.
Retention keeps the newest fulls and removes every incremental chained to a superseded one. The state is made durable before anything is deleted, and a restart reconciles leftovers.
Where you reach it
- CLI
kynetra backup schedule --store <dir> --state <file> --enc-key-id <id> --enc-key-hex <hex> [--rpo-secs --rto-secs --full-every-secs --keep-full --keep-incremental] [--once] - CLI
kynetra backup restore-to-seq --store <dir> --target-seq <n> --out <new dir> --enc-key-id <id> --enc-key-hex <hex> - CLI
kynetra backup drill --store --state --scratch --evidence --enc-key-id --enc-key-hex; kynetra backup rpo-status --store --state
Limits
- The chain store is a local directory. No network call is made, and chain slices are not pushed to an object store by these commands.
- The chain captures the whole instance log, not one project, and every slice is encrypted. Keys are supplied as command-line values, so pass them from a secrets manager.
- It reads the log from position zero, so it cannot be used on an instance whose log has been trimmed.
- Cross-region replication helpers, an immutable vault with retention locks and object-lock header builders exist in the backup crate as library code and are not wired into the command line. Local-directory vault immutability is enforced by the library alone, not by the operating system.
- Run these commands against a stopped server or a copy of the data directory. The scheduler opens the data directory directly.
Evidence
- On this siteReliability
- TestsCovered in the engine repository by chain-integrity and refusal tests (8), scheduler and recovery-point tests (4), retention-ordering tests (6 plus 1), a randomised restore corpus for both entity profiles, 7 drill-scheduling tests, trim refusal tests (3) and 2 CLI exit-code tests
Test coverage is listed as present in the engine repository at this edition. This page does not report test results.
KynetraDB is in private preview
Request a preview to talk through the limits above against your own workload.