archive · sorted desc by ts build log

20 entries on disk · page 1/3. Each post + every image lives inside one embedded OxiDB data dir — no external store, no migrations, no schema files.

0001 · 2026-05-12
skipping the Value tree — 5× faster find→wire decode

@admin 7 words oxidb-embedded

Edit test — applied successfully via curl.
0002 · 2026-05-11
encode 3× faster, output 40% smaller — one-line codec rewrite

@admin 269 words oxidb-embedded

Writes have always felt slower than reads on this engine and I never figured out why. So I wrote a micro-bench that breaks encode apart by document shape and ran it against six payload …
0003 · 2026-05-10
from 4 minutes to 200 ms — fixing create_collection contention

@admin 226 words oxidb-embedded

A scale-bench tenant-creation workload — 1000 collections, eight workers in parallel — hung for four and a half minutes before producing any throughput at all. CPU was idle. The p99 for a single `create_collection` …
0004 · 2026-05-09
ACID hardening — atomic persist, WAL fsync, graceful shutdown

@admin 235 words oxidb-embedded

Durability used to be 'mostly, probably'. Now it's 'ack means on disk', verified by two crash-test harnesses. The persist path was `fs::write` — not atomic, not durable. It got rewritten to: write a `.tmp` …
0005 · 2026-05-08
S3 without the IAM dance — a blob bucket inside the engine

@admin 244 words oxidb-embedded

The blog images on this site are stored exactly the same way an S3 bucket stores objects, except the bucket is a directory inside the OxiDB data dir and the API is a Python …
0006 · 2026-05-07
BM25 + 18 stemmers — ranking that doesn't lie

@admin 265 words oxidb-embedded

Full-text search ranking moved from TF-IDF to BM25. The old scorer told you what words appeared, not which documents were actually relevant. Long documents with high raw term frequency consistently outranked short, focused ones …
0007 · 2026-05-06
OCC transactions across collections — prepare, validate, commit

@admin 286 words oxidb-embedded

Transactions are optimistic and three-phase. No locks held between operations, no deadlocks because there's nothing to deadlock on — readers and writers never block each other until commit time. API: tx_id = db.begin_tx() db.tx_insert(tx_id, …
0008 · 2026-05-05
indexes — single-field, composite, unique, TTL, and when each pays

@admin 318 words oxidb-embedded

Indexes are `BTreeMap<IndexValue, BTreeSet<DocumentId>>` behind a `RwLock`. That structure carries the whole feature set: **Single-field.** Point lookup is `O(log n)` map probe + `O(1)` set head. The bigger win is index-backed sort: a `find` …