49b4bf0177c40800146fb0301c4f9b89d596830d
1335
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 vh
|
ccc0df6870 |
fix(upgrade-docker-ce): retry the stack restart under sudo before reporting FAILED
The restart loop runs as the deploy identity, not root, and a stack .env is allowed to be root-owned 0600. compose bails on the unreadable file before doing anything, so the stack was reported FAILED while restart=unless-stopped had already brought it back healthy — a false failure, which is worse than a quiet one because it trains readers to skim the failure lines. Retry under sudo -n before calling it a failure, and print compose's own output either way. Verified on nh3-dev against beszel: plain attempt rc=1 'open /opt/docker/compose/beszel/.env: permission denied', sudo retry rc=0 'Container beszel-agent Started', container back to healthy. The happy path is unchanged — the sudo attempt only fires after a failure. Also record that tts-dev migrated talk from ~/talk into /opt/docker/compose/talk, which removes the one stack on this host that was invisible to anything walking that path. |
||
|
vh
|
92a4114b90 |
feat(nh3-dev): migrate to docker-ce 29.8 + compose plugin; drop compose v1
Operator cleared the swap and ruled out a docker-compose v1 shim. Ran playbooks/upgrade-docker-ce.yaml: docker.io 20.10.24 -> docker-ce 29.8.0, docker-compose 1.29.2 -> compose plugin v5.5.1, containerd 1.6.20 -> containerd.io 2.3.5, buildx v0.37.1 added. 12 changed, 0 failed, verify 4/4. talk and beszel-agent back healthy on their restart policies. The pre-state was worse than 'old': there was no cli-plugins directory, so 'docker compose' was not a command and exited 0 on a help blurb — a silent no-op that reads as a successful deploy. Records two things the run surfaced. vastblue-u5-pg and its anonymous volume were removed when the old daemon stopped; the playbook has no rm, prune or purge and five other containers survived, so the cause is almost certainly --rm, unprovable now that the record is gone. It was measured beforehand as zero user tables in every database, so nothing was lost. And the playbook's restart loop runs as infra-ops and cannot read a root-owned 0600 stack .env, so it false-FAILs that stack. Also notes that nh3-dev is the only host where /opt/docker/compose is root-owned; the other four are lkraven. Created /opt/docker/compose/talk as lkraven so tts-dev can move talk out of ~/talk. Normalising the parent is left to the operator. |
||
|
vh
|
25a7d05f51 |
feat(nh3-dev): repoint Hermes at gen-large on the LiteLLM gateway
Operator ruled the repoint; Miranda moves off the paid z.ai Coding Plan onto free local compute. model.default gen-large, provider custom, base_url http://10.250.50.70:4000/v1. Verified by a real turn rather than by config: hermes status reports gen-large / Custom endpoint and a completion through /v1/chat/completions returns 660 tokens. The openrouter/nous credit warnings cleared with it. Records the landmine found on the way: CUSTOM_API_KEY and HERMES_CUSTOM_API_KEY are inert for bare provider: custom — they bind only a named custom_providers entry through its key_env. Without model.api_key the request ships the placeholder no-key-required and LiteLLM 401s inside the response body while hermes status still reports a healthy gen-large / Custom endpoint, so status alone cannot verify this change. Also notes that nh3-dev/hermes/api-server-key must not be rotated until SVOS splits its HS256 signing key off the shared value. |
||
|
vh
|
a3e42282ff |
fix(nh3-dev): strip the Hermes api_server tool surface; flag the z.ai model resolution
Set platform_toolsets[api_server] = [] in ~/.hermes/config.yaml. Measured back on /v1/toolsets: 28 rows before and after, enabled 14 -> 0, registered tools 40 -> 0. The row count surviving matters — SVOS's _hermes_roster derives its required-config line from the full report, so narrowing the surface does not blind the diagnostic. terminal, code_execution, file, browser and computer_use are now unregistered rather than merely unreachable. Separately, record an unresolved finding raised by svos-dev: model.default names anthropic/claude-opus-4.6 and model.base_url names openrouter, but provider: auto with only a zai credential resolves Miranda to GLM-5.3 on the paid z.ai Coding Plan. Nothing under ~/.hermes references the LiteLLM gateway. The operator believes Miranda runs on gen-large, which does exist on LiteLLM. Not repointed — the premise is his to confirm. |
||
|
vh
|
d7e69ddaef |
docs(nh3-dev): record the Hermes Agent gateway sidecar on 127.0.0.1:8765
Stood up hermes-gateway.service (user systemd, vendor 'hermes gateway install') for SVOS/Miranda, which cut over from Worldtree to Hermes on 2026-09-11 and aborts at boot without it. Bearer key vaulted as nh3-dev/hermes/api-server-key. Notes the wide default tool surface: with no roster config the gateway registers all 28 toolsets on the api_server platform, terminal and code_execution included. Loopback bind plus the key are the only containment until SVOS installs its svos_miranda plugin and the platform_toolsets / disabled_toolsets roster is applied. |
||
|
vh
|
cd941f7bf9 |
memory: snapshot — fv-ml1 rebalance (cyberprev→sec, gen-small A3B, all seats native 262K)
Pre-clear snapshot. Rewrote Current state / in-flight to the post-rebalance topology (cyberprev is the sec seat, gen-small stood up, char/coder retuned, all generative seats native 262K in-band, catalog+bench shipped) and added a Recent decisions index entry + detail file 2026-09-14-fv-seat-rebalance-gen-small.md. Also commits two previously-untracked durable artifacts (fv-to-ana-nat runbook + its memory detail) so memory does not lag the work. Archival: file is over the ~300-line soft cap (377) but 0 entries archived — the 4 age-eligible (>14d) entries all carry open deferred-work pointers (guarded), and all other entries are <14 days old. Reported per the archival discipline. |
||
|
vh
|
dfa91a8eaf |
docs(fv-ml1): add curated LLM seat catalog (lineage/provenance/cards/speed) + bench script
Adds docs/pfi/llm-seat-catalog.md, the durable hand-curated record of what each seat IS -- lineage, provenance, model-card facts, quantization, speculative decoding, licenses, and measured warm tok/s + deep-prefill depth results with their harness and date. It complements the auto-generated fv-ml1-gpu-seat-inventory.md (live placement/KV/concurrency): the two split by volatility, and the catalog defers to the inventory for any live number. Adds scripts/seat-bench.py so the catalog's speed/depth numbers are reproducible (warm decode tok/s, n=3, conc=1, median; deep prefill at ~0.97x max-model-len with an allocator-log OOM scan). Serial by design -- concurrent deep prefills would confound both OOM and tok/s. Captures the 2026-09-14 measurements: all six generative seats prefill to ~255K (coder ~16K) with zero OOM/CUBLAS/illegal-memory; warm decode 62.7-337.3 tok/s; per-seat VRAM. seat-inventory.py now cross-links the catalog in its footer. |
||
|
vh
|
05f3d3b2a3 |
feat(fv-ml1): add gen-small A3B seat; sec/gen/char to native 262K; coder reclaim; gateway
Full seat rebalance across GPU0/GPU1 (flash on GPU2 and empty GPU3 untouched), operator-directed. Every target seat now serves native 262,144 context with concurrency in the requested 1.2-2.5x band, verified from live boot logs: cyberprev (sec) 262144 @ 1.37x depth-probed CLEAN to 259,722 tokens flash-next (gen) 262144 @ 1.31x (untouched, already in band) gen-small (NEW) 262144 @ 2.56x MTP k=3 measured 69.6% accept / 3.09 len char-rp 262144 @ 1.22x (was 1.02x; util 0.40->0.52) char-rp-fast 262144 @ 2.04x (util cap 0.30->0.24, pinned KV unchanged) - gen-small: NEW seat, Qwen3.6-35B-A3B (3B active MoE), llmfan46 Heretic (MPOA) NVFP4 experts-only, already on disk at qwen36-35b-a3b-heretic-nvfp4. There is no general Qwen3.8 A3B (3.8 MoEs are Flash-Next and the 2.4T), so this is the 3.6 fallback the operator specified. GPU0, :8026, MTP k=3, coherent and MTP-verified before wiring. gen-small / gen-small-reasoning gateway aliases. - coder: 8192 @ 13.75x -> 16384 @ 4.70x (util 0.06->0.055). Context doubled, waste cut. Not the exact 2-3x target: the 1.5B weight+overhead floor (~4.2 GiB) sits just under the util knob's resolution, so hitting <=3x reliably needs a --kv-cache-memory byte pin (compose change) rather than the util fraction. - cyberprev raised 163840 -> 262144: depth-probed with non-repeating prompts to 259,722 tokens, clean (no OOM, memory flat). Unlike mog-sec (same base arch, capped at 163840 for depth crashes), this checkpoint holds native depth. - Gateway (operator calls): summarizer + classifier -> gen-small; new classifier-large -> gen-large (flash) for the accuracy tier; summarizer-large stays on flash. All verified end-to-end. - GPU1 hit its ceiling raising char-rp; resolved by trimming char-rp-fast's reservation cap (its KV is pinned, so concurrency held at 2.04x) rather than moving a utility seat -- the shared GPU_ID on reward/embed/rerank made a single-seat move messier than the in-GPU rebalance. Seat inventory regenerated from the live containers. |
||
|
vh
|
ff8dfed498 |
docs(fv-ml1): regenerate seat inventory — cyberprev in the sec seat, mog-sec gone
Live-read regeneration after promoting cyberprev into sec/sec-reasoning and retiring mog-sec. GPU0 now carries cyberprev alone; sec/sec-reasoning aliases resolve to :8025. |
||
|
vh
|
207798de66 |
feat(litellm): promote cyberprev into the sec / sec-reasoning seats; retire mog-sec
Operator instruction: take down the existing sec seat (mog-sec) and promote hotdogs (cyberprev) into the sec and sec-reasoning gateway seats. - mog-sec container (vllm-mog-sec, :8019, fv-ml1 GPU0) taken down; ~48 GB freed on GPU0 (cyberprev, already co-resident there, is now the sole GPU0 chat seat). - Gateway sec -> hosted_vllm/cyberprev-27b @ :8025; sec-reasoning -> hosted_vllm/cyberprev-27b-thinking @ :8025. sec/sec-reasoning are ROLE aliases, so this is a promotion, not silent substitution (samplers were already identical between the sec blocks and cyberprev, so only model+api_base changed). - Removed the standalone cyberprev-27b / cyberprev-reasoning gateway aliases added in the prior commit -- now redundant with sec/sec-reasoning, and the fleet convention is a role alias on the gateway with the model's served-name only at the vLLM layer (as mog-sec had). cyberprev's vLLM served-names are unchanged. - Verified e2e through the gateway: sec answers (nmap -sV version detection), sec-reasoning answers with a thinking split (127 reasoning tokens); retired mog-sec-27b now 400s. Note: mog-sec was the fleet's only offense+defense/blue-team seat; the sec role is now offense-only (cyberprev tool-calling). Operator-directed after reviewing the capability comparison. mog-sec stack files retained for a future restore. |
||
|
vh
|
3906c6842c |
docs: correct sec-seat lineage — M.O.G.-SEC/mog-sec is an offense+defense SFT finetune, not a persona-on-stock
The sentinel-r3 header and two memory notes described mog-sec (Blackfrost
M.O.G.-SEC / Qwentium) as 'a persona system prompt on stock weights'. Its card
is explicit that it is NOT: base_model_relation: finetune on Qwen/Qwen3.8-27B,
a refusal-free offense+defense cybersecurity SFT with YaRN 1M context ('not a
system-prompt sticker on a stock Qwen'). So all three sec-seat candidates are
Qwen3.8-27B SFT finetunes and differ in training focus, not in kind:
mog-sec = broad offense+defense SFT; sentinel-r3 = pentest agent-trajectory SFT;
cyberprev = cyber tool-calling LoRA SFT on an abliterated base.
|
||
|
vh
|
196416f3da |
docs(cyberprev-seat): correct model lineage — it is a cyber SFT finetune, not an abliteration
Earlier commits and the compose header called cyberprev "an abliteration of Qwen3.8-27B (refusal-direction weight edit, no finetune)". That was wrong. Per the hotdogs model card (base_model_relation: finetune), cyberprev is an offensive-security agentic tool-calling LoRA, merged at scale 1.0, trained on hotdogs/cyber-sft-agent-qwen38 (8,400 rows across 22 pentest tools: nmap, sqlmap, metasploit, hydra, crackmapexec, ...). The abliteration is INHERITED from its base (hotdogs/Qwen3.8-27B-abliterated); the new capability in this repo is the cyber SFT itself. So the sec-seat lineage is: mog-sec = persona on stock weights; sentinel-r3 = pentest-trajectory SFT on stock weights; cyberprev = cyber tool-calling SFT on an abliterated base. All three are SFT-or-prompt variants; the earlier framing of cyberprev as "not a finetune" was backwards. Corrects the compose header lineage block, the homepage dashboard description, and the gateway alias comment. No functional change; the running seat and its weights are unchanged (label refreshes on next container recreate). |
||
|
vh
|
b94d173887 |
fix(cyberprev-seat): serve coherent v2 quant from name-repaired source; wire gateway
The first cyberprev quant served pure gibberish despite passing every structural check. Root cause: the hotdogs/Qwen3.8-27B-abliterated-cyber-preview checkpoint ships broken -- 850 of 1199 tensors (the whole transformer body) carry a TRIPLED `language_model.` prefix, confirmed present in the upstream HF index (an unsloth export bug, not our download). transformers silently loaded the unmapped tensors as random-init, the pipeline quantized noise, and structural verification passed because it inspects the output config, never that input names mapped. A plain greedy coherence smoke caught it; nothing structural could. Fix: rewrite tensor names against a known-good reference of the same architecture (qwen38-27b-uncensored-bf16), verified the repaired name-set == reference name-set exactly (1199 tensors, 0 collisions), and re-quantized from the corrected source (cyberprev-bf16-fixed). v2 is coherence-verified (spec off AND live with dflash): Paris/Berlin/Rome, correct arithmetic precedence, primes, clean cyber-domain chat with a proper thinking split. dflash k=7 measured at 2.77 mean acceptance length. Seat serves cyberprev-27b / cyberprev-27b-thinking on :8025/GPU0 with dflash k=7, displacing the retired sentinel-r3. Gateway: removed the retired sentinel-r3 / sentinel-r3-reasoning aliases and added cyberprev-27b / cyberprev-reasoning -- NOT repointed, since serving cyberprev weights under sentinel's name is silent substitution. Verified end-to-end through the gateway; sentinel-r3 now 400s. Playbook gains a superseded-claims row: structural verification does not imply a good quant; a coherence smoke is the load-bearing gate, and third-party sources must have their tensor names diffed against a reference before quantizing. Broken v1 quant dir removed. |
||
|
vh
|
1418edb03e |
feat(cyberprev-seat): mixed-NVFP4 quant of the abliterated cyber-preview, displacing sentinel-r3
Third sec-seat candidate: hotdogs/Qwen3.8-27B-abliterated-cyber-preview, an abliteration (refusal-direction weight edit) of Qwen3.8-27B aimed at the cyber-offense refusal surface -- distinct from mog-sec (persona on stock weights) and sentinel-r3 (SFT finetune). Operator instruction: quant it, take sentinel down, serve it with mtp or dflash. Quantized to the house mixed recipe via services/gen-seat-mixed-quant/ (NVFP4 W4A4 on MLP layers 0-55 + FP8 W8A8 on attn/linear_attn/lm_head/MLP 56-63, FP8 KV). The prior attempt (2026-09-11/14) died with "Cannot determine num_attention_heads" because it ran from a bare .venv whose newer compressed-tensors reads that field at top level; quant_mixed_nvfp4.py already promotes text_config attention fields for exactly this reason, and the run through the canonical vLLM-image + llmcompressor 0.13.0 / compressed-tensors 0.18.0 path (versions recorded from the container) completed clean. post_quant.py did its job: grafted the 15 MTP tensors verbatim (BF16), and re-injected re:^mtp.* into the ignore list after llm-compressor pruned it for matching no module at quant time -- without which vLLM loads the grafted head uninitialised and speculative decoding runs at 0% acceptance. It also caught a missing preprocessor_config.json (absent from the abliterated source AND its hotdogs upstream); restored from Qwen/Qwen3.8-27B, verified byte-identical to the working sentinel-r3 build, so the vision tower preprocesses. Verified from the tensors, not the config: NVFP4 covers MLP 0-55, FP8 covers 56-63, no overlap, 168 weight_packed tensors (56x3), 15 BF16 MTP, 333 BF16 vision. 51.0 GiB bf16 -> 21.0 GiB. Served under its own name (cyberprev-27b / -thinking), NOT sentinel-r3 -- serving different weights under a retired name is silent substitution. Takes over :8025 and GPU 0 (co-resident with mog-sec). dflash k=7 is the default, carried from the sentinel measurement, but is re-measured against MTP on this ABLITERATED body before cutover, since abliteration is exactly what can desync an MTP head. Drafter (qwen38-27b-dflash2-drafter) shared with mog-sec. Context capped at 163840 not native 262K, inherited from mog-sec's hard-won depth lesson; raise only after a non-repeating deep-prefill probe on this checkpoint. |
||
|
vh
|
4954cf6756 |
test(flash-next-seat): depth-probe orca to 258K clean; record a blind instrument
Bisected context depth on the orcarouter checkpoint with non-repeating prompts (unique random hex per probe, so prefix caching cannot short-circuit the prefill). Six depths from 31,978 to 258,517 tokens, all served. The load-bearing evidence is the engine allocator log: zero OOM, CUBLAS, or illegal-memory entries across the run. That is the same detector that caught the dealignai near-miss at 155K on the previous checkpoint, where it did fire. The probe also ran under real concurrent operator load, making it a stricter test than a solo run rather than a weaker one. Positive control passed: a mis-sized first attempt produced a ~265K-token prompt and got a clean 400 naming the limit instead of killing the engine, so the probe could detect the failure mode it was looking for. Calibration for re-runs: random hex words tokenize at 7.9 tokens/word here. vLLM #54919 (long prefill starving decode for 3-7 minutes) did not reproduce: 258K prefilled in 28.9 s, roughly 8,900 tok/s, scaling near-linearly from 32K. Records that the probe's memory-headroom half was BLIND and must not be reused. It reported an identical 95,460 MiB used / 2,427 MiB free on every row across an 8x range of depths, which is the tell. Two causes: --kv-cache-memory pins the pool and the engine logs "skipped memory profiling", so GPU usage is flat with respect to depth; and the actual risk is a transient activation spike during prefill, which before/after nvidia-smi bracketing structurally cannot observe. Peak-activation headroom therefore remains unmeasured; the pass/fail result rests on the allocator log alone. Also qualifies the earlier 167.5 tok/s decode figure as a possibly-contended lower bound, and records the operator's independent 140 tok/s average measured in real use while this probe was loading the same card. |
||
|
vh
|
d6cc22f6a0 |
chore(flash-next-seat): hardlink duplicated orca shards — 76 GiB reclaimed
The PLE conversion copied 28 non-PLE shards instead of linking them, because os.link hit EXDEV across two separate container bind mounts of the same filesystem. Both directories sit directly on /tank, so the duplication was recoverable. All 28 files verified byte-identical by SHA-256 before any change -- the point of a hardlink is that two paths become one file, so "same size" is not sufficient evidence. Relinked with ln to a temp name followed by an atomic rename over the target, never rm-then-ln, which would leave a window where the file does not exist for a process that reopens it. Generated as 56 literal-path commands rather than a loop, per the no-destructive-ops-on-a-variable-path rule. 294G apparent across the two directories -> 218G actual. Done with the seat live and serving; it did not blink. Records the coupling this creates: the pristine and converted directories now SHARE INODES, so editing a shared file in place in either one changes both. config.json and model.safetensors.index.json are deliberately excluded from the sharing, since the conversion changed them. Also consolidates two duplicate disk bullets left by earlier edits. |
||
|
vh
|
1b5d6ba23a |
docs(flash-next-seat): dealignai weights deleted — record that no local rollback exists
Operator instruction: delete the displaced dealignai checkpoint. 125 GiB reclaimed from /tank (59% -> 57% used). Verified before removing: not mounted by any running or exited container, no symlinks, no inodes shared with the converted orcarouter directory. Every "rollback is two .env keys" statement across the stack README, the .env.example, persistent-memory and its detail file was true when written and is false now -- the .env backup still names paths that no longer exist. Corrected in place rather than left as false reassurance, since a stale rollback instruction is discovered precisely when it is needed. Reverting this seat now costs a 126 GiB re-download. The quality A/B against dealignai is likewise no longer runnable locally: its reference arm is gone. The pristine 170 GiB orcarouter download is retained deliberately -- it is what makes the PLE bf16->FP8 conversion reproducible without re-fetching -- and that is now recorded so a future session does not reclaim it as an obvious duplicate. Also notes that ~75 GiB of non-PLE shards are duplicated between the pristine and converted orca directories (the convert's hardlinks hit EXDEV across two container bind mounts); both now sit directly on /tank, so relinking would reclaim it if /tank ever tightens. |
||
|
vh
|
4390be947d |
feat(flash-next-seat): serve orcarouter weight-only NVFP4 on gen-large
Swaps gen-large from the dealignai ModelOpt W4A4 build to orcarouter/Qwen3.8-Flash-Next-Uncensored-NVFP4, which is weight-only on both axes (W8 float attn, W4 float experts, input_activations: null) and so avoids the 4-bit-activation long-context degradation mode. The checkpoint was previously recorded as unloadable on any mainline vLLM, requiring a from-source PLE-loader patch. That conclusion was wrong on cost. Qwen4ExpPLEEmbeddingMethod.from_quant_config checks ple_embedding_dtype as branch 1, before any quant-config type check, and its NotImplementedError for CompressedTensorsConfig is scoped to the PLE path only -- experts and dense load through the ordinary compressed-tensors paths. Verified by instantiating the real config and calling the selector both ways before doing any work. orcarouter ships a bf16 PLE, so the fix was to make the declaration true: convert the 51.2B-param table to FP8 and declare it. Its 128 PLE tensors sit in one shard file with nothing else in it. Global amax 0.0894, per-shard outlier ratio 1.66x, scale chosen exactly representable in bf16 so no scale-rounding error stacks on quantization; amax maps to 446.17/448, no clipping. Round-trip 2.655% RMS relative, 0.002% underflow, 0 saturation -- the same FP8-PLE treatment dealignai already shipped. MTP head (31 tensors) and vision tower carried through untouched. A second, independent blocker followed: orcarouter labels its 12 QSA layers qwen_sparse_attention, which vLLM rejects; it accepts full_attention and selects QSA via indexer_n_heads. Confirmed indexer_n_heads == 4 in both this and the dealignai checkpoint before renaming -- without that check the rename silently selects plain attention and serves a subtly wrong model that still passes a healthcheck. Measured on the live seat: healthy, coherent, KV 344,155 tokens @ 262,144 ctx, MTP k=3 at 60.4% acceptance / 2.81 mean acceptance length, warm decode median 167.5 tok/s at conc=1 (n=5, spread 12.2%). The reorg note's dealignai figure came from a different harness, so this is not claimed as a win over it; what it does establish is that weight-only experts did not cost decode speed. Still open: controlled quality A/B vs dealignai, and a deep-prefill probe at 262K. Rollback is two .env keys; dealignai remains on disk. Also corrects the README's MTP-is-off section, stale since k=3 was deployed, and adds a superseded-claims row to the quantization playbook. |
||
|
vh
|
ed3f1ede3b |
memory: snapshot — FV recovered, fv-ml1 seat reorg, gen-large blocked on orca runtime
FV colo recovered 2026-09-13 midday (chassis on PDU, firewall on the Eaton 5P1000, GPU caps 275W/card). All-night fv-ml1 seat reorganization: - flash-next gained MTP k=3 (campaign measured it a win here, +52% at conc=1), inverting vLLM's 4xH100 recipe; KV 14->10 GiB. - gen consolidated onto flash-next (all 8 gen/summarizer/classifier/judge aliases repointed); 27B dense gen seat retired, 38 GB freed on GPU0. - char-rp restored to the in-house MeroMero-v2-31B dense heretic (was serving a leftover-test RedHatAI 26B); char-rp-fast is the deliberate speed tier. - Sentinel-R3 (SFT pentest finetune) served for an A/B vs mog-sec, then dflash k=7 cut over after measuring it beat MTP (2.40 vs 2.18 acceptance, ~121 tok/s warm). gen-large is intentionally DOWN: the orcarouter weight-only NVFP4 build downloaded (170 GB, verified) but no mainline vLLM loads its compressed-tensors qwen4_exp PLE; the third-party backport was vetted and is unfit (old-hardware fork, no Blackwell image). Runtime decision pending -- this is the resume point. Also this session: vh/infra-reference repo, scripts/seat-inventory.py + daily drift alarm, OPNsense API reference vendored, secrets shed from a prior scratchpad. Leaves the fv-to-ana-nat files (another session's) and graphify-out untouched. |
||
|
vh
|
c2a70c13db |
feat(fv-ml1): daily drift alarm for the seat inventory
Wires scripts/seat-inventory.py --check to a user systemd timer on nh3-dev (09:15 daily, Persistent=true so a missed run fires on next boot) that posts to althing when the committed document stops matching the live box. Alarms rather than auto-committing. A drift means something changed on the HOST, which deserves a human look -- silently regenerating the doc would erase the evidence of when the change happened and why, which is how the char-rp substitution went unnoticed for three weeks. The alarm includes the changed table rows, not just the fact of divergence, so it does not send the reader hunting. ⚠ The post goes --to infra-ops, which is the fleet ops handle the reading session also runs as. That is the documented exception -- a memo from cron to a future session, the same pattern as the Beszel alerts -- so the message says so in its first line, to stop a future session triaging its own alarm as peer mail and trying to reply to it. SuccessExitStatus=0 1 because a detected drift is a deliberate non-zero exit, not a unit failure. |
||
|
vh
|
a91b841d86 |
feat(fv-ml1): generate the seat inventory from the live box instead of maintaining it by hand
The seat documentation must stay current, and a hand-written document cannot. The LiteLLM config described char-rp as a 31B model on a host and GPU it had not been on since 2026-08-24 -- three weeks of silent drift in a file that read as authoritative, and the reason a seat spent that period serving a model nobody intended. Anything typed here drifts the same way; anything read off the running containers cannot. scripts/seat-inventory.py derives the whole document from the host: - placement and VRAM from nvidia-smi compute-apps, mapped to containers through /proc/<pid>/cgroup -- nvidia-smi reports the vLLM engine child while docker reports the container pid, so matching them directly silently yields nothing - weights and KV tokens parsed from each engine's own startup log, not derived arithmetically, with concurrency computed as KV tokens over context - architecture, layer and expert counts, and the exact quantization group scheme (W4A4 vs W4A16 distinguished) from each model's config.json - speculative-decoding method and k from the container argv, which is how the three incompatible methods on this box became visible - lineage from the .PROVENANCE.txt SIBLING files -- they sit beside the model directory, not inside it, which is why an earlier pass wrongly reported two fully-documented seats as having no provenance - gateway aliases resolved from the LiteLLM config on ana-docker --check compares the committed document against the live box and exits non-zero when they diverge, ignoring only the generation timestamp. Suitable for CI or a scheduled drift alarm; read-only throughout, safe against production. Also commits the KV_CACHE_BYTES override added to the MTP campaign runner, which asserts the flag exists in the derived argv and aborts rather than running a campaign that silently ignored it. |
||
|
vh
|
2d83a895c1 |
docs(fv-ml1): reconcile GPU seat inventory and model lineage against running containers
Read from docker inspect, nvidia-smi and each model's own config.json rather than from compose files or the gateway config, so the document records what is deployed rather than what was intended. Findings worth naming: - char-rp drift. The LiteLLM config documents char-rp as G4-MeroMero-v2-31B on ana-ml2 GPU 0; :8016 actually serves gemma4-26b-a4b-it-nvfp4, a 26B 30-layer 128-expert stock Gemma-4, on fv-ml1 GPU 1. Host renaming explains part of it, but 31B MeroMero-v2 and 26B stock gemma4 are different models and the comment block carries sampler defaults keyed to the model it names. - Two gateway aliases are dead: erp-tune-v2 (:8098) and gemma4-26b-a4b-it-base (:8099). Nothing is listening on either; callers get a connection failure rather than a clear retirement error. - Three different speculative-decoding methods coexist -- dflash k=7 with a separate drafter on mog-sec, qwen3_5_mtp k=3 on gen, mtp k=3 on flash-next. The method string is architecture-specific; copying a --speculative-config between seats without changing it will not work. - Three seats run the floating :latest tag, so what they would pull on recreate is not what is running and was never recorded. - Seven aliases resolve to vllm-gen on :8015, which is why GPU 0 shows draw during ordinary OpenWebUI use -- its auto-titling fires summarizer per turn. Also commits the flash-next-seat MTP k=3 change made today: compose gains the speculative-config with its measured rationale, and .env.example documents the KV-cache reduction MTP requires (14 GiB OOMs, 10 GiB verified in production). |
||
|
vh
|
f9a8f176ca |
fix(mtp-bench): give each arm its own container name instead of tuning teardown waits
Container teardown latency here is unpredictable, not merely slow. Measured 2026-09-13 on the same container in the same session: once ~55s, once 0s. Removing a container holding ~92 GB of GPU memory plus the offloaded PLE mapping leaves an Exited record owning its name for that whole window, so any fixed wait or retry budget is a guess that will eventually be wrong -- a 12x5s retry lost arm k1 by roughly two attempts. Arms now use $NAME_BASE-$arm, set in boot() after stop_bench has torn down the previous arm. Names are never reused, so the collision is impossible by construction and teardown latency stops being load-bearing. cleanup() reaps every fn-mtp-bench-* container at exit. stop_bench keeps the GPU-memory wait -- the next container genuinely needs the device, and unlike the container record nvidia-smi clears promptly and reports truthfully. It no longer waits on the container listing at all. Three earlier hypotheses were wrong and are recorded so they are not retried: the name is not held by a phantom (the container is present and Exited), the removal does not fail (rm -f succeeds; it is asynchronous), and GPU memory release does not track name release (memory frees first, by a wide margin). |
||
|
vh
|
aa5ee7e0a6 |
fix(mtp-bench): docker ps -a is the wrong probe for name release; wait on GPU memory and retry the conflict
The previous fix polled `docker ps -a` until the bench container's name disappeared. That probe is useless here and the fix was a no-op: measured 2026-09-13, the container stops being listed while the daemon still holds the name reservation, so the poll returns false early and `docker run` hits a Conflict naming a container ID that `docker inspect` already reports as 'no such object'. Arm k1 was lost twice this way. Two changes, each aimed at something actually observed: - stop_bench waits on GPU memory falling below 1000 MiB rather than on the container listing. That is the resource the next container genuinely needs, and nvidia-smi reports it truthfully. - boot() retries `docker run` while stderr matches 'already in use', up to 12 attempts at 5s. The daemon's name lag is transient, and forfeiting an arm over it is silent data loss -- run_arm turns a failed boot into a skipped arm that still lets the campaign report success. Both probes verified against real artifacts before deploy: grep -c 'already in use' on the captured k1.runerr returns 1, and the nvidia-smi query returns a bare integer that compares correctly under [ -gt ]. The earlier fix passed bash -n and was still inert, so syntax is not the check that matters here. |
||
|
vh
|
67a7dc539e |
fix(mtp-bench): stop_bench raced docker name release, silently dropping arms
`docker rm -f` returns once removal is INITIATED, not once complete. The
bench container holds ~92 GB of GPU memory plus the offloaded PLE mapping, so
the name stays reserved for several seconds while the device is released. The
next `docker run --name` then dies with a name Conflict.
That failure was near-invisible: run_arm treats a failed boot as
`arm SKIPPED`, the campaign continues, and it still prints CAMPAIGN DONE.
Observed 2026-09-13 -- arm k1 lost the race after off_A's loaded container,
while k2 won it only because k1 had never started a container to tear down.
Every arm that follows one which actually ran is exposed, so k3 and off_B were
both on track to vanish from a run that would have reported success.
stop_bench now polls until the name is released (120s ceiling, warns and
continues). Added a completeness gate: the campaign asserts every expected
res-<arm>-rep<n>.json exists and reports CAMPAIGN INCOMPLETE naming each
missing result rather than DONE, so a gutted run cannot look like a clean one.
Verified the docker ps format string empirically -- an earlier draft nested
quotes so the template rendered as '{{.Names}}' with literal quotes, which
grep -qx could never match, making the wait a no-op that passed bash -n.
|
||
|
vh
|
0a9cf58e19 |
fix(mtp-bench): peak-power summary mis-keyed GPU 0 by timestamp
power.log lines begin '<HH:MM:SS> 0, <W> W, ...', so splitting the first pipe-field on ', ' yields a[1] = '20:18:00 0' rather than '0'. GPU 0 was therefore keyed by sample timestamp, emitting one bogus row per sample and no recoverable peak, while GPUs 1-3 aggregated correctly. Take the GPU index as the last whitespace-separated token of a[1]. Verified against a two-sample fixture: the old parser emits a spurious row per timestamp, the patched one reports 0/1/2/3 once each at the correct peaks. The end-of-campaign summary is the GPU-side input to the fv-ml1 circuit budget, so a silently wrong GPU 0 row is a measurement fault, not cosmetic. The in-flight campaign still runs the old on-host copy (editing a running bash script corrupts execution by byte offset); its summary will be recomputed from the raw power.log. |
||
|
vh
|
06a68f3f6e |
memory: raise the handoff staleness window from 1 h to 8 h (operator)
A one-hour window deleted the handoff unread across any overnight gap -- which is precisely the case the handoff exists for. Tonight's snapshot would have been discarded before the operator returned from the site visit. 8 h also matches the global CLAUDE.md and the /snapshot skill default, so the repo stops being the outlier. The reasoning is kept inline in the header rather than only in this commit message, because the next session reads the header and not the log. |
||
|
vh
|
816b8d4893 |
memory: snapshot — FV dark, Flash-Next seat built, power policy settled
Operator-invoked before heading to the Fountain Valley site visit. Rewrites Current state / in-flight around the outage: FV dark since 06:56Z including the BMC, will not self-recover, 19 of 30 gateway aliases down with no local fallback because every free local model lived on fv-ml1 and irv-ml1 runs no chat seat. The section now POINTS AT docs/runbooks/fv-site-dark-20260913.md rather than restating it, since the runbook is the single source of truth for the visit. Records what was built and verified before the power failed -- the Flash-Next seat on one card with its 51B n-gram table in host RAM, gen-large on the gateway -- and what is decided but unexecuted: the 250 W / 200 W power caps, the incomplete MTP campaign with one off_A rep banked, and the ten stale Homepage labels that the staged bring-up fixes as a side effect. Closes a stale in-flight claim: the 'fv-ml1 cannot initiate to fleet LAN IPs' gap is resolved, fixed by another session's scoped SNAT at 06:22Z with Beszel 18/18 verified. Adds ana-ml3 as its own subsection, since conflating it with fv-ml1 caused two rounds of wrong arithmetic today. Auto-archival fired (index was 431 lines, over the ~300 cap) but reached only 2 entries. The 14-day guard holds almost everything: this repo has had an exceptionally dense fortnight and nearly every dated entry postdates 2026-08-30. Of the six entries old enough, four carry open deferred-work pointers and were held back per the decision-loss guard. The file stays over cap, which is the documented trade -- an over-cap file that keeps live decisions beats a scannable one that lost one. |
||
|
vh
|
5a5f5c267e |
power: RETRACT the DCGM caveat — config management and power limits ARE supported on our cards
I claimed in
|
||
|
vh
|
100670eed1 |
power: what DCGM is, and why not to plan around it on workstation-SKU cards
DCGM is NVIDIA's own Data Center GPU Manager -- first-party, Apache-2.0, packaged as datacenter-gpu-manager -- and it layers above NVML rather than beside it: nvidia-smi is a thin CLI over NVML's per-GPU primitives, and DCGM is a daemon plus dcgmi adding health, diagnostics, config enforcement, policy and group abstractions on top. Which is why its group notion still resolves to N per-GPU writes underneath. The caveat that matters, and it undercuts the experiment suggested in the previous commit: DCGM is datacenter-oriented and parts of it are gated to datacenter SKUs of the Tesla/A100/H100 class. Our cards are professional/workstation parts -- RTX PRO 6000 Blackwell Max-Q and RTX 6000 Ada -- and several DCGM capabilities are unsupported or degraded outside that line, plausibly including config enforcement, which is precisely the power path. So DCGM_CONFIG_POWER_BUDGET_GROUP may return 'unsupported on this device'. Downgraded from 'worth testing' to five minutes of curiosity after the real work, and explicitly not a planning assumption. None of which touches the plan: nvidia-smi -pl 250 is plain NVML and works on these cards. DCGM would only have bought the group-budget experiment and nicer telemetry, and is probably not installed anyway since beszel-agent-nvidia shells out to nvidia-smi. |
||
|
vh
|
94fb7b7208 |
power: answer the bank-budget question — DCGM has the concept, the dynamic part is a control loop, and 4x250 already is 1000 W
DCGM_CONFIG_POWER_BUDGET_GROUP ('the power budget for the entire group') exists
alongside DCGM_CONFIG_POWER_CAP_INDIVIDUAL, so the concept is first-class. The docs do
not state how a group budget is distributed, and the deduction is that it cannot be
anything exotic: the only enforcement primitive underneath is NVML's per-GPU
nvmlDeviceSetPowerManagementLimit and there is no bank-level register, so any group
budget resolves to N per-GPU writes. Static even division is one write each; 'each card
free until they are all loaded' requires continuous re-writing, which is a control loop
rather than a hardware feature. Worth a ten-minute test when the box returns, in case
NVIDIA already runs that loop.
Records the design constraint that matters more than the logic: power readings lag and
-pl application takes tens of milliseconds, so a reactive daemon overshoots during a load
ramp -- and the ramp is the dangerous moment, being the same all-cards-at-once shape as
this box's ten restart:unless-stopped containers starting together. So any such loop must
be safe-by-default and opportunistic upward: boot at budget/N, only ever raise after
observing idle neighbours. Inverted, it works for weeks and then fails on precisely the
event it existed to prevent.
And the reason to defer it: 4x250 W is already 1000 W, so the static cap is the
conservative floor of the dynamic scheme rather than an alternative. The daemon's entire
contribution is the one-card-busy case, worth perhaps 5% throughput, which is rare for a
serving fleet that puts one seat per card and common only for a training window.
|
||
|
vh
|
b538fde6f0 |
caps: fv-ml1 250W / ana-ml3 200W — and nvidia-smi -pl caps BOARD power, not wall power
Operator set fv-ml1 at 250 W per card (83% of TGP, ~5% throughput) and ana-ml3 at 200 W (67%, ~10-15%). Records the term that decides whether 250 W actually clears a 15 A feed, because it is easy to drop: a power limit bounds BOARD power, and the wall sees that divided by PSU efficiency. Four cards at 250 W is 1000 W of board; add 180-300 W of host components and divide by ~0.90 and the plug sees ~1310-1445 W, against a 15 A circuit's 1440 W NEC continuous derating -- an inference box serving all day being a continuous load. So 250 W lands ON the limit rather than under it, where 200 W would give ~1090-1220 W with real margin. The deciding term is the host draw, which is still an estimate, so the procedure is: set 250 W, verify at the plug under four-card load, fall back to 200 W if it reads near 1440 W. A cap is a claim; the ammeter is the verification. Two consequences recorded alongside. Caps bound sustained draw and not transients -- the enforcement window is short but not instantaneous -- and while a breaker's thermal-magnetic curve forgives brief overload, a UPS's overload protection does not. So 250 W implicitly commits the fv-ml1 chassis to the PDU rather than behind the 1500 VA unit, which it exceeds even capped. And ana-ml3's 200 W across only two cards is deliberately conservative at 400 W total, relaxable if Anaheim's measured headroom beats its trip history. |
||
|
vh
|
2da0c76d99 |
correct the hardware: fv-ml1 is 4x Blackwell Max-Q 300W, ana-ml3 is 2x Ada RTX 6000 — and four cards is a breaker problem
Operator clarification, and it separates two boxes I had been conflating. fv-ml1 is 4x Blackwell RTX PRO 6000 Max-Q at 300 W each (Max-Q being the reduced-TGP SKU; the Workstation Edition is the 600 W part), 391 GB VRAM, deployed and currently dark. ana-ml3 is 2x Ada Generation RTX 6000 at 300 W, 96 GB VRAM, not yet deployed. The 200 W cap directive is ana-ml3's. With the TGP known, the outage stops being a vague 'undersized' and acquires a mechanism: two Max-Q cards at 300 W is ~600 W of card, plus a host carrying 566 GB of RAM, drives, fans and PSU conversion loss at perhaps 200-350 W, against an Eaton 1500 VA's real ~900-1200 W. That lands at or just over the rating, which is precisely what explains a full day of service on one card and failure minutes into the second. The host term is the only one being guessed; idle-at-the-plug measures it directly. It also surfaces something that is not a UPS question at all. Four cards at 300 W plus ~300 W of host is ~1500 W against a 15 A circuit's 1440 W continuous derating, so four cards uncapped is marginal on the breaker with no UPS in the path. Capping therefore belongs at fv-ml1 as well as ana-ml3, or fv-ml1 needs a 20 A feed -- and worth noting today's incident only ever had two of the four cards working. ana-ml3's placement constraints sharpen too: sm_89 has native FP8 but no NVFP4, so the in-house NVFP4 quants stay at FV, and at 96 GB total it cannot host the Flash-Next seat at all -- that needs 74 GiB resident on a single card, and the offload moves the n-gram table rather than the experts. |
||
|
vh
|
8fcc26e2c9 |
policy(gpu-power): cards are RTX 6000 Ada at 300 W — 200 W is a mild cap, plus two sm_89 placement consequences
Corrects the SKU: RTX 6000 Ada, 300 W, not the ~600 W initially recalled. That makes 200 W a cap to 67% of TGP -- the favourable part of the concave perf/watt curve, roughly 10-15% of throughput -- rather than the severe 33% cap a 600 W part would have implied, and it very likely sits above the card's enforceable floor, so the check becomes a formality rather than a gate. The protective value is worth stating: four cards at 300 W uncapped is ~1200 W, which is roughly the neighbourhood that overwhelmed a 1500 VA unit at FV with only TWO Blackwell cards drawing. Capping to 800 W makes a repeat of today a non-event. Two consequences that follow from Ada independent of power, and both are placement constraints rather than details. sm_89 has native FP8 but NOT NVFP4, which is Blackwell-only -- so the in-house NVFP4 quants that most of this fleet runs will not be accelerated on that colo's cards, and its seats want FP8 W8A8 builds or the NVFP4 checkpoints stay at FV. And it unparks the triton-backend item, which is a hard no on Ampere because fp8e4nv is unsupported on sm_86 and was explicitly deferred to Ada; sm_89 has what it needs. VRAM is 4x48 = 192 GB against fv-ml1's 391 GB, so big-model placement stays at FV. The Flash-Next seat needs 74 GiB resident on one card and would not fit a 48 GB Ada card even with the n-gram table offloaded -- the offload moves the table, not the experts. |
||
|
vh
|
3e61d7d4e0 |
policy: cap GPU power limits at build time — 200 W for the other colo's cards
Operator directive, and the right generalisation of the FV outage: decide the power envelope first and size the cards into it, rather than installing cards and discovering the constraint by tripping it. Four cards at 200 W is 800 W, which fits a real circuit with a real UPS and headroom. Records three things to settle before it is a plan. First, 200 W may sit below the card's enforceable floor: nvidia-smi -pl is bounded by Min Power Limit, often around half of TGP on a high-TGP part, and a sub-floor request is refused -- quietly, depending on how it is scripted. Run nvidia-smi -q -d POWER before any build planning depends on the number. Second, the 600 W figure wants confirming against the actual SKU. The Ada parts do not land there -- RTX 6000 Ada is 300 W, L40/L40S 300/350 W, 4090 450 W -- while 600 W is Blackwell RTX PRO 6000 Workstation territory, so these may be Blackwell or the figure may be a two-card total. Read it off the device rather than a spec sheet. Third, the workload asymmetry is in this fleet's favour: decode is memory-bandwidth-bound and tolerates a cap far better than training does, with a concave perf/watt curve where 60-70% of TGP costs roughly 10-15% of throughput. A cap to a third of TGP is deeper into the steep region; measure it on the first card rather than predicting, and expect prefill-heavy and training work to pay more than a serving seat. And persist the cap. A hand-set limit holds until the next reboot and then silently stops holding, which is the worst shape available given that the thing rebooting the box is likely to be the power event the cap existed to prevent. |
||
|
vh
|
b6335bf6ad |
runbook(fv-outage): the circuit case — split power survives a trip on battery, but only if the colo handoff does
Operator: 'unless of course the thing trips the circuit anyway.' Correct, and it splits into two halves with different answers. A breaker trip is the event the split-power proposal survives: firewall + BMC is 25-40 W on a 1500 VA unit, which is hours of battery, and on a trip the UPS stops being a load-bearing supply and goes back to being what it is for. What it does NOT cover is the colo's own handoff -- their switch, ONT or demarc. If that sits on the circuit we just tripped, the outcome is a firewall running on battery with nothing upstream to talk to and the drive happens anyway. Added as a question for the facility, because it decides whether split power delivers remote diagnosis or merely feels like it does. Records the case where none of it matters: removing an undersized UPS does not remove the constraint, it promotes the next one -- UPS ~900-1200 W to circuit ~1800 W at 15 A or ~2400 W at 20 A. Which side the four-card figure lands on decides everything, which is what makes that single ammeter reading the load-bearing measurement of the visit. Surfaces the lever that may avoid an electrician entirely: nvidia-smi -pl caps per-card TGP, so the box can be made to fit its feed at a throughput cost rather than a rewiring cost. Read nvidia-smi -q -d POWER for the enforced range before assuming how much room the dial has, and persist any cap -- one that evaporates on reboot will hold right up until the next power event and then silently stop holding. |
||
|
vh
|
00b842bb9b |
runbook(fv-outage): operator ruling — undersized UPS; NAT demoted; ammeter protocol for the visit
Operator's reasoning, accepted and better than the hypothesis-space argument it replaces: the NAT change went effective, was verified bidirectional, and then ran correctly for twenty minutes before the site died the moment GPU load was applied. A working config change does not spontaneously fail under an unrelated physical variable. The load correlation is tight; the NAT correlation is merely adjacent in time. Undersized UPS is the only candidate that explains the trigger. NAT material retained as record, and the power.log/uptime check demoted from decision point to free confirmation. Adds the measurement protocol, since the operator is bringing a PDU and an ammeter. The load-bearing caveat: power.log is GPU-ONLY -- nvidia-smi per-card, excluding CPU, 566 GB of RAM, drives, fans and PSU conversion losses -- so the ammeter at the plug is the primary instrument and power.log only cross-checks the GPU share. Four states to capture (idle, one card, two cards, four cards), and capture PEAK rather than average: UPS overload protection responds to short-term overload, so an average-only reading that hides transients will mis-size the replacement exactly the way the present unit got mis-sized, and must be recorded as a floor rather than as the draw. The four-card figure is earmarked for servers/fv-ml1/README.md, because it closes the cutover's own open question -- that the FV circuit was likely specced against half the real draw, back when every record still said the box had two GPUs. |
||
|
vh
|
59ddedd980 |
runbook(fv-outage): a NAT change 34 min earlier means power is not established — and power.log settles it for free
Another session applied a scoped Tailscale SNAT rule to the FV gateway at ~06:22Z, 34 minutes before the site went dark (docs/runbooks/fv-to-ana-nat.md, not my work, left uncommitted). That makes the UPS-overload theory a hypothesis rather than a finding, and nobody should buy hardware on it until the discriminator below has been read. On the evidence that change is the wrong shape to have caused this, and it is recorded as such so the visit is not wasted chasing it: one OUTBOUND SNAT rule scoped to a single source /32 and a single destination /16 cannot stop the gateway, the BMC or the public WAN address from answering inbound; no routes, filter rules, WAN settings or subnet advertisements were touched; pfctl -sr came back byte-identical; and it was verified bidirectional afterwards including ANA->FV SSH with Beszel 18/18 up. Their BMC datapoint used 10.251.50.50, which is not the BMC -- that is 10.251.250.50, a different subnet. They correctly declined to claim BMC health, but the observation is void rather than negative and should not be reasoned from. The discriminator costs nothing and is already on disk: power.log is written locally to /tank every 10 s by a shell loop on the box and does not depend on the network. Entries past 06:56Z mean the machine never lost power, which makes this a routing fault and the UPS innocent; entries stopping at 06:56Z confirm power. Cross-check with uptime and journalctl --list-boots -- continuous uptime across 06:56Z kills the UPS theory outright. So the first action on site is now to READ, not to fix. The two hypotheses lead to completely different remediations and only one of them needs a new UPS. |
||
|
vh
|
524da702ef |
flash-next-mtp-bench: run on GPU 3 / :8023 without touching the production seat
Operator-directed so the campaign could run in parallel with live gen-large testing. Strips the production-restore tail entirely -- the only container the driver can create or remove is fn-mtp-bench, and the sole remaining compose call is a read-only config --format json to derive argv. Adds per-card power+memory sampling every 10 s to power.log, because two cards under load was the risk being accepted and a record beats an argument. That power log is now the only surviving measurement of what the load drew before Fountain Valley went dark; it lives on /tank rather than in the container. |
||
|
vh
|
312725ddfb |
memory: snapshot — Flash-Next seat on one card, and the FV outage that followed
Durable capture so tomorrow's session does not have to reconstruct either half. Built and verified before the power failed: Qwen3.8-Flash-Next serving on a single RTX PRO 6000 with its 51B n-gram table pinned in host RAM and read over CUDA UVA -- 74.36 GiB weights resident, 14.00 GiB KV for 560,654 tokens at the full 262,144 context, 67 GiB host RSS -- plus a gen-large gateway alias verified end to end. The five findings worth carrying: the offload is #54371 (UVA, merged) which supersedes the paused worker-based #53899 and designs out its entire bug family; text_config.ple_embedding_dtype is the load-or-fail discriminator for any community build; --kv-cache-memory makes vLLM SKIP memory profiling and ignore gpu-memory-utilization, which inverts the usual pin-bytes advice and let a 16 GiB pin nearly OOM with no visible failure; MTP is off pending measurement here rather than written off, because the recipe's number is cross-harness and tested k=3 only while the head is one layer run autoregressively; and a container once reported (healthy) with no published port at all, because the healthcheck runs inside the boundary it was trusted to validate. Then the outage. Records it as will-not-self-recover, so no session wastes effort polling a dead site, and carries the three things that change the visit: bypass the UPS rather than using its surge-only bank (both banks share one 12 A inlet -- the surge bank bypasses the inverter, not the current rating), recover power.log before anything else because it is the only load measurement that exists anywhere, and bring seats up one at a time because ten restart:unless-stopped containers loading at once is the largest transient the box can make into whatever just failed. Also records what is still half-done: the stale homepage labels on the 10 containers that died before they could be recreated, which the staged bring-up fixes as a side effect, and the eight drifted stacks plus three untracked host-only stacks that were deliberately left for a deliberate reconciliation. |
||
|
vh
|
d79f10457a |
runbook(fv-outage): UPS overload as leading hypothesis, site-visit bring-list, no-local-fallback correction
Operator's read is that the UPS the box was plugged into overloaded and died, and it fits better than the breaker-trip theory: a UPS's output rating sits far below the circuit's, so it is the first protective device to give -- which explains why the site let go at TWO cards loaded rather than four, and why the ~25 W firewall died with it. Records the operationally important consequence: a tripped UPS resets, an overloaded one can kill its output stage permanently. If it is dead, nothing on site can be reset back to life, so the visit needs the means to BYPASS the UPS or it is wasted. Elevates recovery of /tank/.../power.log to the first action on site. It sampled all four cards every 10 s up to the cut, lives on /tank rather than in a container, and is the only measurement of what the load actually drew -- without it a replacement UPS gets sized by guesswork. Also states that no load figure exists yet, only idle. Corrects an earlier claim of mine in this session: there is NO local fallback for the 19 dark aliases. Probed -- every free local model is on fv-ml1, and irv-ml1 runs no chat seat at all, only TTS/ComfyUI/arbo/clipper work on two partly-occupied Ampere cards. The only non-fv chat backends are paid. Any paid coverage must go under a new opt-in alias name rather than a silent repoint of summarizer/gen/classifier. |
||
|
vh
|
969a1b64a2 |
runbook: FV site dark 2026-09-13 — outage facts, blast radius, staged recovery, OOB design gap
Written while the site is down so recovery does not have to be reconstructed later. Records what was measured rather than what is suspected: every FV address including the BMC is unreachable while all three other sites answer, the campaign's last log line was off_A rep 2 at 06:56:04Z, and the site was dark by 06:58:40Z. Names three candidate causes with the evidence that would distinguish them, because the instrument that could have settled it -- the per-card power log -- died with the box. The two-card-load hypothesis fits the timing and the two prior Anaheim breaker trips on this same chassis, but it is circumstantial and is recorded as such. Carries the recovery hazard that matters: every seat on the box is restart:unless-stopped, so resetting power alone brings ten vLLM containers up loading at once -- the largest transient the box can produce, into a circuit that may have just tripped. Staged sequence given, gen first and flash-next last. Also records the OOB gap the outage exposes: OPNsense-as-subnet-router protects against box-down/gateway-up, and not at all against the site-wide loss that actually happened, because the BMC's only path out is through that same gateway. |
||
|
vh
|
7e62a07341 |
flash-next-seat: full 262K context, KV pinned at a measured 14 GiB, gen-large on the gateway
Operator-directed: raise context to the model's native maximum and take as much KV as the card safely allows, and expose the seat through LiteLLM as `gen-large`. max_model_len 131,072 -> 262,144 KV cache 8.76 -> 14.00 GiB (332,721 -> 560,654 tokens) concurrency 2.54x@128K -> 2.14x@262K ⚠ 16.00 GiB WAS TRIED FIRST AND IS TOO AGGRESSIVE. A 155,497-token non-repeating prefill drove GPU 2 to 97,074 of 97,887 MiB and the caching allocator logged "OOM on device 0 while trying to allocate 488636416 bytes (free: 422117376)" -- 466 MiB wanted against 403 MiB free. The request completed, so nothing failed visibly; that is one step before the shape that crashed stacks/mog-sec twice on 2026-09-10 (~1.04 GiB wanted, ~600 MB free). Backed off to 14.00 GiB, which re-probes clean: zero allocator warnings, a 155,557-token prefill in 14.2 s, and 2,085 MiB still free at peak. The reason the first estimate was wrong is worth keeping, because it is not obvious and it inverts the usual advice: --kv-cache-memory makes vLLM SKIP MEMORY PROFILING ENTIRELY and ignore --gpu-memory-utilization. The profiler was the thing accounting for deep-prefill activation, so pinning bytes switched off the protection that the pin was supposed to formalise. vLLM's own "--kv-cache-memory=18745235968 (17.46 GiB) to fully utilize gpu memory" line is computed from a profile measured at max-num-batched-tokens depth and sits 3.5 GiB above what a 150K-token request survives; open #54764 compounds it, since PLE short-conv prefill pads every request in a batch to the batch-MAX query length. max-num-batched-tokens stays at 8192 -- it is what bounds the activation peak, and doubling max_model_len left the profiled peak unchanged at 1.65 GiB precisely because the peak tracks chunk size, not context length. Gateway: `gen-large` added to the LiteLLM model_list, pointing at fv-ml1:8022. One alias on purpose -- a single alias cannot trip the shared-config enable_thinking mutation footgun, which needs two over the same (model, api_base). Sampling is the checkpoint's own declared set (temp 1.0 / top_p 0.95 / top_k 20); presence_penalty, min_p and repetition_penalty are left unset because the checkpoint declares no canonical value for them. Verified registered for both the infra-ops admin key and the shared all-agents key, since a new model behind a scoped allowlist 403s silently. Also adds services/flash-next-mtp-bench/ -- the MTP measurement campaign and its rationale. MTP stays off, but on "not yet measured here" rather than on vLLM's 4xH100 recipe number, which is a cross-harness comparison and not evidence about a TP=1 Blackwell seat. |
||
|
vh
|
3132a16ca0 |
fv-ml1: finish the renumber the cutover missed -- 16 dead dashboard links
Every fv-ml1 link on the Homepage dashboard was broken. Measured against the
live dashboard API before the fix: 16 entries pointing at the dead 10.250.50.54
and zero at the live 10.251.50.54, covering gen, M.O.G.-SEC, Scriberr, Embed,
Rerank, Reward, Coder, Dockge and six dormant seats.
The miss was structural, not careless. fv-ml1-rename-sweep.sh works from an
allowlist assembled from files that mention the HOST, and a homepage.href label
mentions only an IP -- so every stack whose sole stale reference was a label
fell outside it. The allowlist now covers those 24 files, and records how to
derive the list next time (grep the old address, subtract history) rather than
enumerating from memory.
History is still untouched, and the exclusions are now written down with the
reason each one keeps the old address: recorded benchmark results, whose
base_url is part of a measurement's provenance; the one LiteLLM comment
preserving a retired hand-test endpoint; and the cutover runbooks, where the old
address is the subject matter.
Two bugs found while applying it, both fixed here:
- deploy-stack.sh rejected any stack name containing a dot, so qwen3.5-122b,
qwopus3.5-122b and mistral-medium-3.5 could not be deployed by the script at
all. The check exists to stop path traversal, which means rejecting ".." and
"/" -- not every dot. Traversal is now rejected explicitly and tested.
- stacks/scriberr/.env.example allowed CORS only from the dead IP and from
scriberr.ana.internal, which no longer resolves; the box is at the fv site
and DNS already carries scriberr.fv.internal. The live .env had both stale
origins, i.e. an allowlist with nothing reachable in it.
Host side, applied separately: canonical pushed for the 16 stacks whose only
difference from the host was this renumber, and an in-place address-only fix for
the nine whose host copy has genuinely drifted or has no canonical copy, so that
drift survives for a deliberate reconciliation instead of being clobbered. Every
compose.yaml on fv-ml1 now reads 10.251.50.54. The labels themselves only take
effect at container creation, so the running containers still need recreating.
|
||
|
vh
|
f964a47262 |
flash-next-seat: Qwen3.8-Flash-Next on fv-ml1 GPU 2 with the n-gram table in host RAM
First seat whose weights do not fit its card. The model is 176B total -- a 125B main model plus a 51B n-gram (PLE) lookup table -- at ~6B active per token. The table is a pure embedding lookup, so it lives in pinned host RAM and the GPU reads rows directly over CUDA UVA: ~78 GiB resident on a 95.6 GiB card, 47.7 GiB pinned of 566 GB. GPU 2 and GPU 3 were both idle, so this displaced nothing. Checkpoint dealignai/Qwen3.8-Flash-Next-ABLITERATED-NVFP4 @ be794b99, pinned by revision: NVFP4 W4A4 routed experts, FP8 PLE table, everything else at source precision. Chosen over better-liked builds because its provenance states protocols and repeat counts -- AIME26 pass@1 98.75% over 30x8 repeats with a stated SEM, full-set GSM8K, and a byte-equality audit covering all 31 MTP tensors -- and because it declares text_config.ple_embedding_dtype, which is the field vLLM reads first when selecting the PLE weight format. Builds that ship an FP8 table without that declaration resolve to the unquantized path and fail on load; the README records the check. Requires vLLM #54371 (UVA PLE-offload, merged 2026-09-09T14:32Z), verified by ancestry: the pinned nightly is +150 commits / behind_by 0 from the merge commit. Not in v0.29.0, cut six hours earlier. The older worker-based offload (#53899) is paused upstream and is not the path here -- its deadlocks, ptrace gate and stale-output-under-graphs bugs all came from the separate worker process that UVA does not have. Five deliberate departures from the other seats on this box, each from a measurement rather than a preference, all annotated in place: - no MTP: the vLLM recipe measured it worse at every concurrency on 4xH100 (8-36% less throughput, 32-173% more latency, ~36% acceptance) - modelopt_fp4, not compressed-tensors: only the ModelOpt reader honours the ignore list keeping attention, shared experts, PLE and MTP out of W4A4 - KV left at auto: fp8 KV on this model's QSA path is an unmerged RFC (#54426) - mamba-cache-mode stated explicitly: the model raises on mode "all" - 128K context and 8192 batched tokens, not the native 262K: #54764 and #54919 make depth the risky axis, and sizing to the KV pool has never fixed a depth-driven crash on this hardware Nothing is wired into LiteLLM. Pointing an alias at this seat changes what existing callers receive and is a separate decision. |
||
|
vh
|
a5691ce796 |
restic(esh-vm-db): a backup that exited 0 nightly while keeping an April dump
Work by a parallel session on 2026-09-12; committed here with the rest of the day's changes. Rationale in persistent-memory.d/2026-09-12-esh-vm-db-restic-repair.md. The only visible symptom was a systemd-failed unit from a Sep 6 repository network timeout after boot. The real fault was quieter and much worse: the pre-backup hook logged failures as WARN and returned zero, so pg_dumpall could fail every single night -- it used TCP localhost and wanted a password nobody supplied -- while restic dutifully backed up the stale April 23 dump still sitting in the staging directory and reported success. Mongo was fine, which is part of why it went unnoticed. Postgres now dumps over the /var/run/postgresql socket with peer auth and -w, and both database failures now fail the backup rather than masking it, while still preserving any prior per-DB dump rather than truncating to nothing. An ERRORS counter replaces the warn-and-continue path, and the staging directory is overridable via RESTIC_STAGE_DIR so the new test can exercise it. Adds backup.contract.md, retry.conf and test_pre_backup.py -- three red-green regression tests covering the failure modes above. systemd drop-ins on both jobs add network-online ordering plus Restart=on-failure with a 5m delay and a 3-per-hour limit, which addresses the original boot-timeout symptom. Verified against a real run: snapshot bc5eeaff at 07:01 PDT with a fresh 3.46MB PG dump, retrieved from the repository with decompression and completion marker checked (not a full restore). Repository check passed, 99 snapshots. The old hook and stale dump are preserved root-only at /var/lib/restic/repair-20260912. |
||
|
vh
|
3e727dbeb5 |
memory: snapshot — FV cutover executed, BMC online, four GPUs not two
Captures the Fountain Valley cutover as executed rather than planned: fv-ml1 live on 10.251/16 and serving, BMC recovered after finding it was tagging 802.1q VLAN 250 into an untagged port, and the corrected four-GPU / 391 GB count that invalidates prior sizing assumptions. Also records two things that cost real time and would cost it again: OPNsense write APIs need an X-CSRFToken scraped from a script block, not a hidden form input, and a 403 that coincides with the box going down looks exactly like a successful reboot. Auto-archived 8 settled entries (Recent decisions: 8). The index stays over the 300-line cap at 385 because the guards hold -- nearly every remaining dated entry is under 14 days old or carries open deferred work, and an over-cap file that keeps live decisions beats a scannable one that lost any. |
||
|
vh
|
48e34e1849 |
playbooks: repoint the four remaining live fv-ml1 targets
homepage-regroup, mog-sec-move-to-gpu0, pull-hf-repo and serve-qwen3.5-122b all carried runnable 'scripts/elway ana-ml2 --playbook ...' instructions or the old 10.250.50.54 address. Each would fail today against a dead name and a dead IP, so these are corrections rather than cosmetics. homepage-regroup is renamed to match; the other three keep their names, which never carried the host. |
||
|
vh
|
87d7555fcf |
fv-ml1: the playbook + homepage body edits the rename commit missed
A zsh glob failure in the staging step of |
||
|
vh
|
17586abb19 |
homepage: retire the fv-ml1 docker relay, its precondition is now met
stacks/ana-ml2-proxy existed for one reason, stated in its own README: ESH could not reach the box's Docker API cross-site because of a return-route gap, so ana-docker relayed :2354 -> :2375 from the same subnet. The README named its own retirement condition -- "once it has a working return route to ESH, point Homepage back at it directly and compose down this stack". That condition is met. FV is on the mesh, vb-gateway accepts routes, and the NH3 UDM, ESH UDM and Anaheim FortiGate all carry 10.251.0.0/16 statics toward their local scale nodes. Verified before removing anything: esh-docker-vm opens a TCP connection to 10.251.50.54:2375 directly. Homepage now points at 10.251.50.54:2375 and discovers 112 services across 20 groups, so nothing regressed. The socat container is stopped and removed from ana-docker; the compose file and README are deleted here rather than renamed, because renaming would preserve a moving part that no longer earns its place. |