FAQ

Frequently asked questions

Can I run a public Elah Minecraft network today?

No. Elah 0.0.3 is an early read-only inspection release. It does not provide a playable cluster.

Is Elah a Minecraft server fork?

No. The planned runtime uses Minecraft server workers and coordinates them. Elah is not intended to replace the server simulation.

Will players need a mod?

The design goal is to keep normal Java clients. Players should not need a special launcher, client mod, infrastructure resource pack, or custom protocol extension to join.

This remains a design goal until a gameplay milestone proves it.

Does Observer change my world?

Observer opens observed world inputs read-only and contains no production path for repairing or rewriting them. Use a backup or filesystem snapshot anyway. Good backups remain important for every world administration tool.

Can Observer scan a running server?

It can detect many changes and stop, but it cannot create a consistent live snapshot. Stop the server or use a filesystem snapshot.

Does Observer tell me when a player last played?

No. It reports saved timestamps only as metadata. Those timestamps are not proof of player activity.

Does Elah already split chunks between machines?

No. Multi-worker ownership is future work.

Why not start with automatic scaling?

Automatic scaling depends on safe ownership, handoff, storage, recovery, and fencing. Building those foundations first makes corruption and duplication less likely.

Is Elah Kubernetes for Minecraft?

It is a useful comparison, not a literal description. Both can have a control plane, scheduler, placement constraints, health evidence, and elastic capacity. Kubernetes manages generic workloads and machines. Elah is intended to make decisions from Minecraft topology, ownership, MSPT, TPS, players, entities, chunks, handoffs, generation, and storage.

Kubernetes may eventually run workers underneath Elah, but it remains optional.

Can several cheap machines always replace one fast machine?

No. Spread-out survival, exploration, many independent bases, and generation work can be easier to divide. One dense lobby, mob farm, combat arena, or other tightly connected hotspot may still depend on strong local simulation performance.

Elah must measure whether work can move or split before requesting more capacity.

Could a host share one fleet between customers?

That is a future integration goal. Every customer cluster would need separate controller authority, credentials, storage, networks, metrics, ownership state, and hard resource limits. Billing and customer management would remain outside Elah.

What happens when machines disagree?

The planned system fails closed. It should prefer temporary unavailability over allowing two authoritative writers for the same area.

Where should developers look?

Start with the technical guide, then use the repository's contributing guide for development procedures.