---
title: Storage
description: Eviction savings, dev-store throughput, repair, and SQL bounds over upstream denoland/celld.
---

These patches change what Cello stores and how it reads it back. For bucket request counts, read [efficiency](/project/what-cello-adds/efficiency); for the durability contract, [what Cello guarantees](/concepts/durability). Each entry names its fork patch number.

## Evict without redundant bucket reads

- `0054` — A bucket-posture eviction skips visibility LISTs when this residency itself uploaded into the prefix: a successor restores by listing, so the uploader's own success response is proof enough. A resumed local copy, recovered tails, bundles, and the L0-chain fallback still list.
- `0056` — Concurrent bucket-posture writes on one cell share ownership reads: a gated write joins a read in flight on its cell and epoch when the issuing gate's cover reaches its position, and issues its own read at once otherwise. No write waits longer than before; fleet proofs never share reads.
- `0057` — An eviction deletes the per-cell L0 and L1 objects its handoff snapshot supersedes: after the snapshot PUT succeeds, a background task deletes this epoch's journaled upload keys (max txid at or below the snapshot's) with one single-key `DeleteObject` per key — never bulk `POST ?delete` — at bounded concurrency. The eviction returns at once; failed DELETEs leave restore-invisible garbage for epoch GC. Journal capped at 8192 entries per cell epoch.
- `0061` — Follow-up: the eviction guard proves a journaled epoch with one HEAD on the newest journaled key instead of a prefix LIST, falling back to the LIST on a miss, an error, or an empty journal.

## Develop against a faster local store

See [local development](/guides/local-dev) for the workflow:

- `0099` — The dev store funnels every PUT through one writer thread, one connection, and one queued batch per transaction. Concurrent PUTs no longer lose the SQLite lock lottery into `busy_timeout` sleeps: an idle batch costs one commit like today's single PUT, a loaded batch shares one fsync, and each row keeps its own savepoint so create/update conflicts still fail alone ([upstream #310](https://github.com/denoland/celld/issues/310)). Local dev-store only; fleet and bucket durability paths untouched.
- `0122` — Local LISTs read from a covering index over a prefix range instead of scanning the whole `objects.sqlite3`: the open creates `objects_listing` and bounds the scan to the prefix, with the Rust-side prefix check as backstop. No new knob; dev-store only ([upstream #265](https://github.com/denoland/celld/issues/265)).

## Repair crashes and bound resources

- `0098` — Compaction retains live-named epoch files: after an L1 fold the node keeps the newest covered L0 file that `pos()` and verify still read back, and each activation plus each startup drops live epoch dirs a newer local epoch supersedes ([upstream #311](https://github.com/denoland/celld/issues/311)).
- `0100` — The 0.5.1 facet-image import is repaired: the check is no-user-data instead of `page_count == 0`, which never fired under replication, and the decision is stamped in `user_version` so a later wipe never resurrects it ([upstream #292](https://github.com/denoland/celld/issues/292)).
- `0105` — Unclean restarts prune live-named epochs; see [fleet and durability](/project/what-cello-adds/fleet-durability) ([upstream #290](https://github.com/denoland/celld/issues/290)).
- `0066` — A per-connection SQLite page-cache budget: `finish_open` sets `PRAGMA cache_size=-<kb>` from `CELLO_SQLITE_CACHE_KB` (default 2000, the stock ceiling: no behavior change), validated at startup. See [configuration](/reference/configuration).
- `0102` — Hydration counting is O(1): the hydration set carries an exact cut-relative count maintained on every insert and truncation, so a faulted read no longer walks the whole cut ([upstream #270](https://github.com/denoland/celld/issues/270)).
- `0128` — Per-statement SQL time is bounded by `CELLO_SQL_STATEMENT_MS`: when set, each statement is interrupted past the bound and fails with a catchable timeout error that rolls the statement back; unset means unbounded, the historical behavior. Streaming cursors share their call's single deadline, and a timeout flag from an armed run never rewrites a later unarmed error ([upstream #274](https://github.com/denoland/celld/issues/274)). See [configuration](/reference/configuration).
