Docs / Operations Operate
Backup & Recovery
Back up the workspace, and recover from torn writes and stale indexes.
Backup
The ledger is append-only, so a consistent copy is straightforward:
- Stop the indexer (or wait for it to finish).
- Copy
.ekos/(andekos.toml).
For fact-engine ledgers, sealed segments never change; only the active segment and the search index are mutable, so incremental copies of sealed segments are safe. Artifacts and caches (.ekos/artifacts, llm-cache, embed-cache) are rebuildable and can be excluded — the LLM cache is worth keeping if you want to avoid repeating paid calls.
Recovery
| Situation | Action |
|---|---|
| Crash during a write (torn active-segment tail) | ekos ledger repair — verifies sealed segments, heals the tail |
| "search index schema is stale" from a read-only open | ekos ledger repair (or any writable command), then reopen |
| Search results missing after a failed commit | ekos ledger repair |
Lost .ekos/ entirely |
Rebuild from sources: the pipeline is deterministic, but LLM-derived text and review decisions (events) will not be reproduced unless the LLM cache and ledger were restored |
Review decisions are events in the ledger, so back up the ledger, not just the sources.