feat(nwsync): emit blobs in parallel with a bounded worker pool (#80)
build-binaries / build-binaries (push) Successful in 2m21s

Closes #79.

Emit is latency-bound, not CPU-bound. Every blob costs two serial round-trips to the zone — a `ProbeKey` HEAD, then a `PutReader` PUT — so a hak with a few thousand resources pays a few thousand serialised latencies. Measured on the live sow-assets-manifest backfill: 26 s of CPU across 9.5 minutes of wall clock, on a 4-core host with 5 GB free and peak RSS of 51 MB.

`emit` now hashes, compresses and stores `--jobs N` resources at once, default 16 — matching `DEPOT_JOBS` and the transport's `MaxIdleConnsPerHost`, so a worker per connection needs no fresh TLS handshake. `--jobs 1` is exactly the old behaviour.

### Three properties had to survive

Each has a test in `internal/nwsync/jobs_test.go`:

- **Deterministic manifest bytes.** `emitterVersion` promises a manifest is a function of its artifact, so `entries` is index-addressed rather than appended to — a worker owns `entries[i]` alone and the slice comes back in artifact order whatever order uploads finish in. `TestEmitProducesTheSameIndexAtEveryJobCount` diffs the `.nsym` and its sidecar between `-jobs 1` and `-jobs 16`.
- **Index still lands last.** Any worker's failure aborts before a manifest is written. `TestEmitLeavesNoIndexWhenAParallelUploadFails` fails every blob PUT with 16 workers in flight and asserts no `.nsym` appears. Under `-race` it also covers the shared counters.
- **Identical content still shares one blob.** This one bit during development and is the reason to read the diff carefully: serially, the sink's existence check absorbed two resrefs with identical bytes. In parallel both workers probe, both miss, and both upload — `TestEmitWritesBlobsAndManifest` caught it as "wrote 2 blobs, want 1". Claiming the sha1 in-process restores the dedupe and skips a probe round-trip as well.

### Memory

Peak now tracks the resources in flight rather than one resource. The ceiling is `N` × the 15 MB `fileSizeLimit` plus its compressed copy — bounded by a constant this package enforces itself, and still not tracking the archive. `TestEmitPeakMemoryIsBoundedByJobCount` re-runs the #76 regression check at `-jobs 8`: a hak 8× bigger still costs the same.

This is only cheap because of #78. Before streaming emit, N workers would have meant N whole archives resident.

### Not done

Skipping the `ProbeKey` HEAD on a first-time emit would halve round-trips, but doubles uploaded bytes on a re-run — which is exactly what a backfill is. Noted in #79 so it is not rediscovered; parallelism is the better lever and this PR takes it.

Worker compression still serialises on `blobEncoder`, which is `WithEncoderConcurrency(1)` for the memory reason in `compressedbuf.go`. At 26 s of CPU per hak that is not worth trading memory for, but it is where to look if the numbers ever say otherwise.

🤖 Generated with [Claude Code](https://claude.com/claude-code)Reviewed-on: #80

Co-authored-by: vickydotbat <vickydotbat@tutamail.com>
This commit was merged in pull request #80.
This commit is contained in:
2026-07-31 15:25:33 +00:00
committed by archvillainette
parent fa32dd411f
commit 1c2acc5530
4 changed files with 246 additions and 13 deletions
+10 -1
View File
@@ -48,10 +48,19 @@ beside the artifact itself with the extension replaced, so `emit` and
`assemble` agree on where it is without being told.
```
nwsync emit [--as NAME] [--out DIR] <artifact-key> <file>
nwsync emit [--as NAME] [--out DIR] [--jobs N] <artifact-key> <file>
nwsync assemble --group-id N [--tlk-key KEY] [--out DIR] <artifact-key>...
```
`emit` is latency-bound, not CPU-bound: every blob costs an existence probe
plus an upload, and a measured backfill spent 26 seconds of CPU across 9.5
minutes of wall clock. `--jobs N` (default 16) sets how many resources are in
flight at once. The manifest is byte-identical at any value — the number of
workers is never observable in the output. Peak memory is `N` times the
per-resource limit of 15 MB plus its compressed copy, so raising `N` far past
the default costs real memory for little gain: the transport keeps 16 idle
connections per host, and past that a worker pays a fresh TLS handshake.
Both verbs upload by default; nothing bulky is ever written to the runner's
disk. `--out DIR` writes a local repository tree instead, which is the
conformance path against upstream `nwn_nwsync_write`. The zone comes from