b193 (U11b) retired the legacy plane and removed --legacy-mode from the
harness; passing it there fails the daily batch. The wrapper now greps
core/memory_acceptance at the deployed sha and omits the flag when it is
gone, so it is correct on both sides of the b192 -> b193 deploy (checked:
7a83f2f -> passes it, c2d87263 -> omits it).
althing-seat-daemon took an outside SIGTERM on 2026-09-27 and exited 0, so
Restart=on-failure never revived it; hermes-gateway sat pull-only for five
days with a yellow verdict nobody was watching. Fixed the death mode with a
Restart=always drop-in (also on the jekyll twin, same latent bug) and added
this watchdog for every other death mode: 5-min timer, alert on the second
consecutive non-200 (~10 min sustained), re-alert at most hourly, recovery
mail on green, exit 2 distinguishes a broken alarm wire from a down seat
(backup-freshness precedent). Lifecycle exercised end-to-end against a dead
port before install.
Prime's ruling 2026-10-02. retention.py picks the snapshots to delete:
the newest 48, plus the newest of each of the last 30 days and of each of
the last 12 ISO weeks, counting only days and weeks that have snapshots.
Names that are not exactly YYYY-MM-DD_HHMM are never selected, and the
NAS side refuses any path outside that pattern. The unit fails unless
the number kept equals the number expected. Live run: deleted 1, 0
errors, 48 kept as expected.
Prime grew VM 102's scsi0 from 250 to 378 GB after the root alert hit 85%
twice in 14 h. The swap partition sat right after sda1 and blocked growth,
so the playbook moves swap to a 4 GB /swapfile, deletes sda5/sda2, grows
sda1 in place (start sector unchanged) and ext4 online, and sets initramfs
RESUME=none so boots don't wait for the vanished swap. Root is 372 GB, 58%.
The in-browser recorder calls getUserMedia, which browsers refuse on http://
origins, so it sat at "Initializing recorder...". scriberr.nh3.phasefinal.com
is now fronted by the fleet TLS caddy on nh3-dev (wildcard cert) and added to
Scriberr's ALLOWED_ORIGINS; the Homepage link points at it. The plain
http://10.251.50.54:8080 URL keeps working except for recording.
Image local/mia:0.1.0 built from stacks/mia: MIA v2 @ bbd8b158 (MIT) with
its pinned submodules, dread-dev's proven Python lock with the torch family
swapped to cu129, and a driver adapted from dread-dev's run_mia.py that
seeds every mesh (fix_random + trimesh's module RNG) and writes
weights_effective into the npz. Weights stay in fv-ml1's shared HF cache
at pinned revisions, mounted read-only.
scripts/mia-run mirrors blender-run: --job DIR is shipped to
fv-ml1:/tank/mia/jobs, one docker run --rm rigs every mesh, out/ comes back.
Acceptance on the four Dread Naught characters: 3.9-4.7 s a mesh (median
of 3) plus 12.7 s model load, peak 3,394 MiB; seeded runs bit-identical
across rotated mesh order; GPU-vs-CPU distances the same size as sampling
noise, with an unseeded GPU run as the positive control.
rsync -a copies a read-only source dir (0555) as read-only, so the hourly
prune's rm -rf could not unlink inside it. From 2026-07-18 every pruned
snapshot was left as a 22-entry husk while the run still logged OK. The
1,740 husks on nh3-nas were removed (0 errors; 48 full snapshots kept).
The prune now runs chmod -R u+w before rm -rf, logs its error count and
the number of snapshots kept, and exits 3 on a failed prune (exit 1 on a
failed rsync) so the systemd unit shows failed instead of passing.
The gen-small EngineCore OOM (04:21 PT): our parked 3,582 MiB window cache left
no room for vLLM's runtime workspace. Seat-side fix, three controls:
- windowed path wraps every window in torch.cuda.empty_cache(), so the seat
returns to ~2,108 MiB rest after a 12-min file instead of parking at the
peak (measured: peak 3,028 MiB during, rest after, restarts=0);
- MEM_CAP_MIB=3840 hard set_per_process_memory_fraction: over-cap requests
answer 503 with the seat alive (proved at cap=2000), so the failure lands
on us, never on a neighbour;
- CUDA_GRAPHS=0: the graph decoder pins cache blocks that empty_cache must
free (illegal-memory-access wedge when both were on first try). Cost:
12-min file 3.0 s vs 1.2 s, short bins 35-62 ms vs 33-42 ms -- still 4-15x
under the sherpa seat.
Measured, not computed: gen-small moved ZERO from 36,116 MiB across three
realistic requests (1,351 in / ~180 out) -- its workspace lands at engine
init; the growth window is restart-relative, matching infra-ops's observation.
WINDOW_S is now a real compose tunable. README memory section rewritten.
Prime-approved switch of the fleet STT seat (fv-ml1 :8300, LiteLLM ext-stt/
whisper-1, caller talk) from the sherpa-onnx int8 seat to arm B-bf16w of the
2026-09-30 A/B (docs/pfi/parakeet-seat-ab-2026-09-30.md): p50 33/36/42/71 ms
vs the old seat's 187/308/626 measured on the same card today, WER 1.965/3.026
vs the A/B floor 1.97/3.09. All three seat defects fixed: 12-min file 200s
(windowed at 360 s after a GPU 0 OOM on one whole request; the A/B's own
long-form method), no pause truncation, no long-form dropout.
GPU 0 room: gen-small --gpu-memory-utilization 0.48 -> 0.36 (0.46 and 0.40
refuse their boot check; cyberprev+voices hold the card). Its KV is byte-
pinned, so the boot log is token-identical: 670,142 tokens / 2.56x before
and after. Seat rests 2,088 MiB; GPU 0 keeps ~1.9 GB free.
Two runtime landmines documented in the README: NeMo's attention mask is
materialised T x T even under local attention (hence the window), and
httptools 0.8.0 writes a NUL into the HTTP status line that httpx — i.e.
LiteLLM — rejects, so the image ships plain uvicorn with --http h11.
Old seat stopped, not removed: docker stop parakeet-nemo && docker start
parakeet is the rollback.
License: NVIDIA Open Model License (accepted by Prime 2026-09-30); note in
stacks/parakeet-nemo/README.md.
A/B of the live STT seat (fv-ml1 GPU 0, sherpa-onnx int8 v3) against
nvidia/parakeet-unified-en-0.6b, measured on GPU 3 with the seat's own image,
k2-fsa's published unified int8 export, fp32/fp16 exports made with k2-fsa's
recipe, v2 int8, and NeMo 3.0.0 (fp32, bf16 autocast, bf16 weights).
- Seat int8 graph runs on one CPU thread (cpu/wall 1.00, GPU 2-9%).
- unified-en under NeMo: -121/-234/-530 ms vs the seat at 1-3/3-8/8-20 s
(paired, n=120/bin; floor <=6 ms; +50 ms positive control reads +52-54).
- unified-en WER lower in every runtime: -0.7 pp clean, -1.5 pp other,
-3.2 to -4.4 pp AMI (paired CIs exclude 0).
- Seat defects found: hard 400 s input ceiling (HTTP 500), truncation after
a quiet 1.5 s pause, and severe long-window dropouts (int8 v3 only).
- B-bf16w needs +0.8 to +1.5 GB over the seat's 1,690 MiB on GPU 0.
Raw requests, hypotheses, manifests and the full harness under
services/parakeet-ab-2026-09-30/. No deploy; live seat untouched apart
from 240 light test requests.
Prime: Scriberr gets the basic fix, v3 stays (no NeMo 3.0.0 surgery). 0002 moves
from proposed/ into the carried set; scriberr-rebuild now applies 0001+0002 by
default (suffix dropout2) and its memory budget becomes a 5,600 MiB regression
guard (Scriberr is on GPU 3). Live on fv-ml1 1602: scripts rewritten from the
patched embed, a 20-min file at 5,502 MiB with retried_gaps reported.
The ~6.5-9 s first-call-per-bucket autotune lived in the container's writable
layer and died on every recreate. 0.1.3 creates /tmp/triton-cache in the image
owned by 10001 so the named volume intern-decision_triton-cache inherits a
writable mount point, and compose mounts it.
Bucket model PROVEN, not inferred: 2,048-token buckets, 16 up to 32,768. After
one warmed call per bucket, 12 random sizes across 8k-32k were all warm (worst
2.09 s); cold entries cost 6.5-9 s. Full cold warm-up 109 s; warm re-run 17 s.
scripts/intern-decision-warmup: one noul call per bucket, MAX_TOKENS from
/health, two-point live calibration of the tokenizer's linear token model (a
single probe overcorrects and the aim oscillates around the bucket edge),
per-bucket wall times, non-zero exit on a missed bucket. Run it after an IMAGE
CHANGE only; the volume carries ordinary recreates (measured: force-recreate,
then a warmed 32k call answered in 2.11 s).
Acceptance on 0.1.3: JevBench 202/231, hard 83/111, 0 diffs / 924; warm 32k GPU
1 peak 15,218 MiB (budget 15,220; a COLD autotune touched 15,224 once, README
caveat); /decide answers. Artifacts in the acceptance dir.
Prime's ask (via the coordinator): investigate the "Parakeet skips
stretches of speech" finding, including other Parakeet weights.
Investigation only; nothing deployed.
Against ground truth (official SCOTUS transcript, Gutenberg #38916) the
drops are real: production v3 loses 140 / 66 clean words per transcript on
the two public files and ~50 on each private one (Whisper-referenced,
Canary-confirmed; adjudicator 129/129 correct on the calibration). Cause:
the v2/v3 0.6B weights collapse deep inside long full-attention windows;
the encoder output is degraded, the audio alone transcribes fine, and
1.1B TDT/RNNT/CTC and CTC-0.6B never do it. Decoding (CUDA graphs, greedy
variants, max_symbols, beam), slice length, local attention, loudness,
resampling and a noise floor do not fix it. Controls: A-vs-A, silence
positive control (>=15 words 36/36), null control, bootstrap floor.
Proposed patch 0002 re-transcribes >=3 s stretches where the audio holds
speech but no word came out (-80 to -90 % lost words on all four
recordings, lower WER, no invented text, +10 MiB) and adds an explicit
PARAKEET_MODEL_PATH with the loaded model recorded in JSON and ModelUsed.
Reviewed at high effort, all findings fixed; built and tested as
scriberr:local-blackwell-a353078-dropout2, not deployed.
scriberr-rebuild: --patches takes DIR[:DIR...]; embeds and seam-checks
both Parakeet scripts (seam-check --standard for the short-audio one).
Prime 2026-09-30: move scriberr to GPU 3 and extend the Jev endpoint to 32k tokens.
Scriberr holds 0 VRAM idle; verified a 20-min job on GPU 3 at 5,496 MiB. With GPU 1
freed, intern-decision's measured card peak at MAX_TOKENS=32768 is 15,220 MiB against
a 15,437 MiB budget (n=3, 1 and 16 questions); 32,769 tokens is refused 422 up front.
JevBench v1.2.16 via /v1/systemone unchanged: 202/231, 0 diffs vs the bench.
Scriberr moved to fv-ml1 GPU 3 on 2026-09-30, so the card is no longer idle when the rebuild runs. The memory stage now requires >= 20 GB free instead of an idle card (GPUs 0-2 still fail that) and attributes the peak only to the host PIDs of its own container, captured with docker top alongside the 0.2 s nvidia-smi samples. Verified with five processes on the card: peak 5,496 MiB, identical to the exclusive-card figure.
Audit finding on 0.1.1: a Jev client that always sends an images array with no
images was rejected for nothing. Only a non-empty value is 422 now. Also aligns
the contract's response example with the wire (model is the name@revision
string, not an object).
Re-accepted live: images []/null -> 200, ["a.png"] -> 422; JevBench all 202/231,
hard 83/111, 0 changed rows across the bench's r1..r4 (924).
Straight passthrough to the checkpoint's own DecisionEngine.predict — never the
semif mapping, whose different prompt would change the answers. Reuses the one
inference thread, bearer auth, MAX_QUEUE, VRAM cap and error envelope; no new
concurrency. 1..16 questions in ONE call (never chunked: Jev questions share a
prompt); images 422; over MAX_TOKENS 422 before the forward. /health advertises
the surface. Response 'model' is a string name@revision (JevBench's runner
hashes it; a dict broke its manifest step).
Acceptance on the live service (see acceptance/systemone-2026-09-30/): JevBench
v1.2.16 typesafe adapter over the 231 public items scores all 202/231, hard
83/111, with 0 changed answers across all 924 rows of the bench's own r1..r4;
controls 401/422x3 (token boundary proven at 7168 pass / 7169 refuse); GPU 1
per-process peak 9,866 MiB under the largest accepted request (budget 9,876);
/decide/shared positive control unchanged. 120 tests green.
vh