# `mix ankusa.recover`
[🔗](https://github.com/jamescarr/ankusa/blob/ankusa-v0.7.0/packages/ankusa/lib/mix/tasks/ankusa.recover.ex#L1)

Brings a stopped node's store forward from the intake log (`wal: :dual`, ADR
0002), by the same code a booting node runs.

    mix ankusa.recover --instance default --node ankusa-a
    mix ankusa.recover --from earliest --dry-run
    mix ankusa.recover --allow-gap

The node's `Ankusa.Config` is `config :ankusa, :config, [...]` (a keyword list
of `Ankusa.Config.new/1` options, `wal: :dual` and its `intake`, `backup` and
`source_store` blocks included): the task needs the node's own configuration,
never a copy of it. Start the replacement node with the lost node's
`intake.node_id` on an empty volume: the store is restored from the backup
first, then the log is replayed from its `m:intake_offset`. A running node does
this at boot; run the task by hand with `intake.recover_on_boot: false`.

Prints `records=… applied=… present=… collided=… unroutable=… skipped=…
republished=…` and exits `0`; a second run changes nothing.

Options:

  * `--instance` — default `default`
  * `--node` — asserts the node id (must equal `intake.node_id`)
  * `--from` — `backup` (the store's own offset, the default), `earliest`, or a position
  * `--dry-run` — read and count, write and publish nothing
  * `--allow-gap` — go on when the log no longer reaches back to the restored
    store's offset, saying which positions are lost

---

*Consult [api-reference.md](api-reference.md) for complete listing*