# `Hyper.Node.Reclaim`
[🔗](https://github.com/harmont-dev/hyper/blob/main/lib/hyper/node/reclaim.ex#L1)

Boot-time reclamation of device-mapper and loop devices orphaned by an unclean
shutdown (SIGKILL or `:erlang.halt`, where the owning GenServers' `terminate/2`
never ran to tear them down).

Hyper names every dm device it creates with a `hyper-` prefix (`hyper-thinpool`,
`hyper-rw-<vm>`, `hyper-img-<img>-<n>`), so this removes exactly those - never an
operator's unrelated dm devices. Removal is leaf-first: a device still open by
another (the pool under a thin volume, a snapshot under the next in its chain)
refuses until its dependents are gone, so leftovers are retried until a pass
removes nothing new. Loop devices backing files under Hyper's data dirs are then
detached (the dm devices that held them are gone by that point).

Entirely best-effort: every failure is logged and boot continues. It runs once,
before any device-owning GenServer starts, so the freshly-booting node never
collides with its own previous instance's leftovers.

# `run`

```elixir
@spec run() :: :ok
```

---

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