From the studio a single connection to Hetzner nbg1 (188 ms away) is
capped near 0.1 MB/s, although a browser speed test shows 50 Mbps.
Measured: 1 connection 0.13 MB/s, 8 -> 0.64, 24 -> 1.69 MB/s. At the old
speed a ProRes MOV could not finish inside the 30 min request timeout.
- Worker 1.7.2 uploads files >= 16 MB as S3 multipart: 8 MB parts on
uploadConcurrency connections (config.json, default 16), each part
retried up to 4 times; aborts the upload if a part keeps failing.
- artifact-presign returns uploadId + per-part presigned URLs when asked
for multipart; new E20b artifact-complete stitches or aborts.
- Old workers and old servers fall back to the single PUT.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Preview jobs looked stuck after aerender finished: the worker was silently
uploading the review MP4s and the ProRes MOV over a slow uplink, and a
restart mid-upload threw the preview away.
- Worker 1.7.1 replaces the job's log tail with live upload status and logs
each upload's size, duration and speed.
- The ProRes MOV now uploads after finalize via a persistent queue
(HighResUploader, resumes after restart, 5 attempts) and attaches through
the new E13b endpoint POST /api/ext/render/jobs/{id}/attach-highres.
A MOV from a superseded preview is discarded.
Deploy the server before the 1.7.1 worker.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Worker 1.7.0 (preview.highQuality, default on) now, after the AE render:
- encodes a High quality MP4 from the delivery MOV (full-res H.264 High,
CRF 14 capped at 100 Mbps) -> Version.hqFileUrl
- uploads the MOV itself -> Version.highResKey/Filename/Size
Both best-effort (the preview never fails over them), with a lease
keep-alive during the long encode/uploads; the export timeline records
what was attached.
Player: HQ switch when a High quality source exists; keeps frame, play
state, speed and zoom, remembered per browser, falls back to Standard if
the file won't load. ProRes itself stays download-only (browser support).
Downloads: the client portal's High Res button serves the version's
ProRes (client-code filename), falling back to the shot's manual file;
staff get a ProRes button on the review page.
Housekeeping (lib/render-pipeline/high-res.ts): HQ MP4 + ProRes are kept
while a version is latest, client-shared, approved or younger than
preview.highResRetentionDays (14); otherwise deleted from storage. Runs
per task after finalize and hourly from the reaper; deleting a version or
a reslate removes its files. The Standard MP4 is never touched.
Also: ProRes/Handles top-bar buttons go icon-only below lg so they don't
overlap the version selector.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two independent features landing together:
Generated shot thumbnails — a shot can now carry a thumbnail promoted
from a render export (Shot.generatedThumbnailUrl/generatedThumbnailFrom)
alongside the existing brief-supplied thumbnailUrl, which always takes
display precedence (lib/shot-thumbnail.ts) so a promoted frame can only
ever fill a gap, never overwrite the brief. Shows with
pipelineConfig.autoShotThumbnail promote automatically on every
finished export (lib/render-pipeline/preview.ts); promoteExportThumbnail/
clearGeneratedThumbnail (actions/shots.ts) are the manual equivalent for
everyone else. Worker-side seek logic for picking the promoted frame is
covered by RenderWorker/VFXReviewWorker.Tests/ThumbnailSeekTests.cs.
Bulk shot delete — the shot-status page's selection bar gets a "Bulk
Actions" dropdown (alongside the existing Change Due Date) with Delete
Shots. bulkDeleteShots() (actions/shots.ts) deletes one at a time rather
than in a single statement: exports_shotId_fkey is ON DELETE RESTRICT
by design, so a shot with renders or deliveries on record is protected
and gets skipped and reported by name rather than aborting the whole
selection or being silently destroyed along with its history.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Production fallout after enabling it: the trimmed EXR footage started
after the head handle as intended, but the overlay's burn-in timecode
(hardcoded to start at frame 1001) did not shift with it, so the
delivered burn-in no longer matched the footage. The picture-lock
reference render also never worked correctly in practice. Reverts the
AE panel checkbox, the preview-build script's trim/reference logic, and
the server/worker plumbing around it back to their pre-feature state
(82dfaee) — untrimmed previews, no picture-lock reference, overlay
timecode correct again.
The noPictureLock/piclockRefKey schema columns and their migration are
deliberately left in place: they're already applied in production, and
dropping an applied migration's file risks the next `prisma migrate
deploy`. They sit unused and harmless until this is revisited.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Previews used to ship the full padded (handle-included) render to
clients, with an artist manually pulling picture lock and hand-trimming
the comp to match. The automated preview stage now trims the client
comp to the picture-lock edit itself (a fixed handle count off each
end, per the studio's existing convention) and separately builds an
ungraded, full-duration picture-lock reference MP4 as a diagnostic
sidecar to check back against. A new "No Picture Lock" panel checkbox
is the single override for shots where picture lock is missing or
wrong, falling back to today's untrimmed behavior.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Blender work is not a shot submission: no Export, no version, no
validation, QC or delivery. It is "Ctrl+F12, but on the farm", so it is
modelled as what it is — a queue entry and nothing more.
That turned out to need very little. Claim, progress, fail and
complete-render already tolerated a job with no Export behind them, so a
Blender job reuses the whole lifecycle; RenderJob gains projectId, label
and submittedBy so a job with no shot code still has something to show.
Capability gating is new and was already missing: the claim now filters
job types by the machine's reported tools. Without it a workstation
without Blender would claim Blender work and fail it repeatedly, burning
the attempts while a machine that could have rendered it sat idle. The
worker asks for Blender jobs unconditionally and the server refuses them
— trusting the worker to self-select would have left the same hole.
The runner treats a clean exit with no saved frames as a failure:
Blender exits 0 when a frame range is empty or an output path is
unwritable, and a job that silently rendered nothing is the worst
outcome available.
Standalone jobs auto-requeue on a transient failure like any other work;
delivery builds are deliberately excluded, since a half-written package
wants a human first.
Adds the Blender add-on (BlenderPanel/) and its docs. The add-on refuses
to submit an unsaved file — the farm opens the .blend from disk, so
unsaved changes would render the previous state without warning. Its
payload is pinned by a contract test, since add-on and server only ever
meet over JSON.
Worker 1.5.0. Adds migration 20260815000000_blender_render_jobs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Initialise Workspace built UNG_101..UNG_117 from a hardcoded range and
imported one hardcoded AEP, and fourteen findComp calls named UNG's comps
literally. None of it worked for a second show.
Comp names now follow a convention derived from the show id —
{showId}_VFX_OVERLAY, {showId}_EXPORT_TEMPLATE, _SHOW LUT — which
reproduces UNG's existing names exactly, so nothing changes for the show
it was written for. A project whose template departs from the pattern can
override any of the three.
Paths cannot be derived: UNG's show id is UNG while its folder is
UNGO_VFX. Those live on Project.pipelineConfig and are edited under
Project → Settings → Pipeline. They are resolved server-side and handed
to the panel with the project list, and read by buildPreviewManifest for
the headless build, so the panel and the farm cannot disagree about which
template — or which show LUT — a shot was built with. That mattered more
than the folder names: a preview rendered through another show's LUT is
quietly the wrong colour rather than obviously broken.
Episode folders come from the project's episode list, falling back to the
folders under its footage root for a show whose episodes are not entered
yet.
Panel 2.5.0. Adds migration 20260814010000_project_pipeline_config.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Some exports are deliberately framed differently from their review —
unscaled or un-punched-in EXRs — and a preview rebuilt from those frames
would be the wrong picture. The previous export's MOV is the correct one.
skipPreview renders the EXRs and goes straight to READY_FOR_QC. Available
at Queue Export (a "No preview" checkbox in the panel, E1 field) and as an
option when overriding a validation failure.
The inheritance is as important as the skip: deliveryMovPath, the review
MP4, the thumbnail and the Version are carried over from the newest
earlier export of that shot which has them. Without it the delivery
builder — which takes its MOV from deliveryMovPath — would package EXRs
with no MOV at all. Where there is nothing to inherit, the export still
reaches QC and the timeline says it has no review media.
QC is untouched by any of this: a skipped preview still needs a human
pass before the export can be delivered.
Panel 2.3.0. Adds migration 20260814000000_export_skip_preview.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
"What is running where" was answerable only by asking someone. Each tool
now reports its own version on startup into a ToolVersion registry, shown
on Pipeline → Machines.
The worker's version moves out of config.json and into the assembly:
a hand-maintained string on each machine is exactly the thing that goes
stale, and a machine reporting a version it is not running is worse than
no version at all. publish.ps1 stamps the git SHA via SourceRevisionId,
so a deployed exe traces back to a commit (1.1.0+feb1c7b).
The panel gains PANEL_VERSION, shown in its title and every diagnostics
report, and pings on load.
Staleness is derived rather than declared: the newest version seen for a
tool is the benchmark, so updating one machine is what flags the rest.
There is no expected-version list to keep in sync — which would have been
one more thing to forget.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Completes RenderPipeline2 phases 3-5. Additive throughout: with
validation.enabled=false the pipeline behaves exactly as before.
Validation (§8) runs on the worker straight after a render, while it
still holds the job and the files are local. Everything is read from the
EXR header — one 64 KB read per sampled frame, no pixel decode — so the
engine needs no oiiotool. The render job deliberately stays open through
the stage, so a worker that dies there is reaped and retried instead of
stranding the export in VALIDATING.
Timecode is the primary check and is anchored on the plate, never the
API: measured on the live show, Shot.timecodeStart is the editorial
in-point and sits one handle-length (8f) ahead of the plate, so checking
against it would pass the defective exports and fail the correct ones.
Reference precedence is manifest timing plate -> plate found on disk ->
API value, and the API rank is WARN-only. The panel now resolves plate
provenance by walking the comp's own footage layers, so the worker never
guesses which element carries the comp's timing.
Verified against real studio renders via the new --validate mode: it
catches both defect classes seen in shipped work — the +8 handle offset
and frame-derived timecode (00:00:41:17 = frame 1001 at 24fps).
QC (§10) is one service behind two front doors, API-key for the panel
and session for the web. PASS means ready to upload and nothing else;
FAIL sets the comp task to CHANGES. The panel gains a QC section that
imports the real EXR sequence at full res rather than the H.264.
Delivery (§11) rides the existing queue as a DELIVERY_BUILD job: hard
links with a copy fallback, verification by re-hashing delivered bytes
against the validation-time digest, manifest.json and the CSV report.
fsutil hardlink list is unavailable over SMB, so link identity is never
queried — reading the bytes back covers links and copies alike.
Also fixes ApiClient.ProgressAsync throwing on transport failure, which
would have aborted work in flight during a brief server outage; progress
reporting is best-effort by contract.
Removes the stale RenderWorker/ rule from the root .gitignore. The
worker's source is tracked like the rest of the repo and its own
.gitignore already covers bin/, obj/ and publish/; the rule silently
excluded new files while the existing 20 stayed tracked.
Not yet applied: migration 20260813000000 was hand-written because the
dev database was unavailable — run prisma migrate deploy and verify
before relying on it. None of this has run against a live server.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A slate or burn-in mistake is usually caught after a preview has already
succeeded, and there was no way back: retryPreview required PREVIEW_FAILED, and
READY_FOR_QC only led on to QC_FAILED/READY_FOR_DELIVERY/SUPERSEDED. The only
recourse was a whole new export, re-rendering every EXR and burning a version
number to fix a typo.
- transitions: READY_FOR_QC and QC_FAILED can now go back to GENERATING_PREVIEW.
- regeneratePreview() rebuilds the manifest from scratch rather than patching
anything, so shot code, episode/scene, frame count, slate date and burn-ins
are all re-derived exactly as on a first-time preview. Validated EXRs are
reused and the version number is untouched -- this re-delivers the same
version, it does not make a new one.
- vfxScope and submissionNote are frozen onto the Export at submission time, so
regenerating refreshes that snapshot from the shot's slate fields (explicit
override wins, same precedence as createExport). Slate text edited on the web
app's Slate tab is therefore picked up.
- finalizePreview replaces the media on the export's existing Version instead of
stacking a second row, so the version number stays put and reviewer notes and
annotations survive the reslate. First-time previews are unaffected.
- POST /api/ext/exports/{id}/regenerate-preview exposes it to the panel.
- Panel: Regenerate Previews button, gated on the same statuses as the server,
with a confirm covering what is and is not reused.
Also fixes the slate description source: the Slate tab authors
shot.slateDescription (falling back to shot.description) but the preview
manifest only ever read shot.description, so edits made there never reached the
rendered slate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Panel export paths now mirror the server's layout
({exportRoot}/{showId}/[{episode} if episodic]/{shotCode}/{version}) via a new
getShotExportDir helper, so farm and local renders land in the same folder.
Folder creation moves into the per-shot loop so one unwritable path no longer
aborts the whole queue.
Also adds an optional `scene` query param to the ext shots endpoint.
One "Queue Export" click now renders the EXR sequence, then rebuilds the shot
headlessly with the studio slate/overlay template to produce the delivery MOV
and review MP4. Implements RenderPipeline2 phases 1-2 plus the preview stage.
Server:
- New models Export, RenderJob, ExportEvent, Machine, WorkerHeartbeat, plus
Project.deliveryConfig and per-submission slate fields (Export.vfxScope,
Export.submissionNote, inherited from the shot's previous export).
Both migrations are purely additive; no existing column is touched.
- lib/render-pipeline: server-enforced state machine, transactional version
increment with supersede, atomic FOR UPDATE SKIP LOCKED claim gated by
machine availability windows, and a lease reaper run from instrumentation.ts.
- /api/ext/* endpoints for the panel and workers; session-auth mirrors under
/api/render and /api/machines for the web UI.
- Pipeline pages: render queue, export detail, machine monitoring, plus an
Exports tab on shot detail.
RenderWorker (.NET 8 Windows service, new):
- Registration, heartbeat as cancel channel, claim loop, aerender runner with
progress parsing and stall watchdog, crash recovery and disk-spooled
reporting that survives server downtime.
- Preview stage: headless AE assembles the preview comp into a throwaway AEP
with both output modules queued, then a single aerender pass renders them.
Preview jobs are not claimed while an interactive AE session is open, so an
artist's project is never taken over.
AE panel: Queue Export with live status polling, urgent flag, retry, and the
VFX Scope / Submission Note fields. Every existing panel action is unchanged.
Preview chaining ships disabled behind SystemConfig preview.enabled.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
twotalesanimation