The rerun after 96c5a06 refused every corpus document again ("document 196 is
8507 tokens and the KV pool holds 7436"), and the boot table shows why the
host-pool change could not have helped: vram_budget.cpp put the container in its
`no lender` branch — elastic 25.56 GiB, experts 25.29 GiB (their plan-fixed
share), kv cache 271.25 MiB — identical at 12,288 and at 49,152 MiB of host
pool. The only supplier left there is the activation arena, and the engine says
so at vram_budget.cpp:117: "lower --max-num-batched-tokens (the arena scales
with it)". The log prices both rungs directly (a 2,048-token step costs 1.08
GiB, a 1,024-token step 504.39 MiB), so the default moves to 1,024: the cache
grows to 866 MiB / 24,300 tokens and scores every one of the 197 documents
(longest 11,970). Calibration keeps its wide 65,536; it was never tied to MNBT.
This also retracts the mechanism 96c5a06 recorded in the header ("the pool buys
KV as well as residency"). Inheriting the env's 49,152 / 96 stays right as
single-source-of-truth, but on this payload the pool is not the cache's supplier.
run-qwen38-bf16.sh carried its own --host-pool-mib 12288 / --gpu-headroom-mib 768,
copied from the w4-era swift15 rig. The engine derives the whole per-card split from
those two numbers, and the KV cache's funding is `elastic - expert_floor` where the
expert floor is a function of the pinned host pool — so the pool buys cache as well
as residency. The first kldref boot got a 270 MiB / 7,436-position pool and the
recorder refused all 197 corpus documents ("document 0 is 10554 tokens and the KV
pool holds 7436"), exiting 1.
The rig now sources FRANK_ARCHON_HOST_POOL_MIB / FRANK_ARCHON_GPU_HEADROOM_MIB from
scripts/frank-archon.env, the file that owns them (49152 / 96), with `:?` guards, and
the header records the law instead of the claim that the pool never touches logits.
tests/qwen38_reference.bats pins the shape verbatim, so its SHAPE contract moves with
it; the suite is green (558 tests).
Two tools the KLD gate depends on.
build-qwen38-bf16-rad.sh converts the checkpoint to a BF16 .rad with NO recipe
line, so every tensor stays bf16 - the reference must not carry the quantization
it is there to measure. It gates before writing 335 GiB: the checkpoint present
and fully verified by the fetcher's own verifier, the peak-disk bill (350,000
MiB: the checkpoint's 343,323 plus margin, because the .rad is built beside it),
an existing output retired to ~/.AGENTTRASH by --force instead of deleted, and
--plan-only before the real convert. --max-ctx 32768 is a declare-time check, not
a weight change: dropped, the resolver is handed the checkpoint's 262,144 window,
which at fp8 cache is 258.8 GiB on a 192 GiB machine.
run-qwen38-bf16.sh serves that container in three modes - kldref (record the
reference logits), score (record an arm's logits against that reference), calib
(open the Gram tap for the imatrix). It drains frank-archon and restores it on
every exit path, and the two logits modes ride ONE step size
(FRANK_QWEN38_MNBT, default 2048) and ONE window (FRANK_QWEN38_MAX_MODEL_LEN,
default 32768) so a reference and an arm cannot be measured at different shapes:
that is the whole validity of the KL number. Four slots over 2,048 matches
frank-archon.env's operating point (MAX_SEQS=4, MAX_BATCHED_TOKENS=2048); the
smaller host pool (12,288 vs 49,152) buys residency, which moves speed, not
logits. calib keeps 65,536 - decode-bound serving, never the gate.
kldref refuses before the drain when the corpus's longest record cannot fit the
declared window (limit = window x 3 chars, the pessimistic tokenizer bound; the
measured corpus is 38,263 chars, 2.6x inside 32,768). It runs on --dry-run too,
so the check is visible before a 20-minute drain is spent. 12 seam tests, each
red-proofed by neutralising the gate it pins.
32 black-box tests over the four tools the campaign runs, driven through their
own argv with fixture corpora - no GPU, no container, no rad-convert. Each test
is an output an operator reads: the verifier's MISSING line and exit, the byte
bill the derivation lands, the refusal text, the plan and the convert the driver
prints per arm, the interval verdict the gate returns.
Includes the planner seam the swift15 campaign lacked: the campaign's paths plan
NO deployable service, so merging this toolchain can never drain frank-archon.
The driver fixture carries four grams so --min-grams has a thin-pool control,
and the arm-driver invocation passes env assignments positionally after the
driver's own flags, so the seam exercises the real call shape.
The swift15 gate compared an arm against the BF16 reference and passed on
KL <= 0.03. That bar is now the wrong instrument: the fleet already ships the
publisher's FP8-native iq4r-moe payload, so an arm that beats BF16 but loses to
what is running is a downgrade. The gate takes three recordings - fp8 anchor,
iq4r anchor, candidate - and requires the candidate's KL strictly inside the
interval the two anchors define, at every slice, plus top-1 agreement no worse
than the served payload.
The anchors must be ordered (fp8 better than iq4r) or the run stops: an
unordered pair means the instrument or the recording is wrong, not that the arm
lost, and one number out of the report is named rather than guessed at.
--recover prints the tightened numbers the same way it prints the failing ones,
so the operator sees the bar they would have had to clear.
One template (every routed expert at five bits, E4M3 scales in 64-groups, the
top level rotated through the 128-wide Hadamard chosen by GPTQ against CALIB)
plus a driver that generates the per-arm expert row, so the other 60 lines can
never drift between arms: control (cd=3), cd=6, and act_order=1 with NO cd= -
libquant refuses the pair (lq_registry.cpp:105), so the AoA arm is its own
sweep, not a fourth column.
The header carries the campaign's order (fetch -> build reference -> one drained
window for the calibration tap and the KLD reference -> trash the bf16 container
-> imatrix -> arms -> fetch the FP8 anchor last -> gate -> trio edit), because
that sequence is point-of-use for whoever runs it next, and the disk sequencing
is forced: 754 GiB free does not hold checkpoint + reference + three arms.
Gates that bite before 5h13m of convert time is spent: the BF16 reference still
occupying the volume, a recording too thin to be a calibration (converting
uncalibrated is silently worse and only whispers on stderr), an imatrix outside
the mounted campaign dir, an arm that already exists, a protected derivation
that cannot land its target, and a destination that cannot hold the output. Each
arm runs --plan-only first so the engine's own fit check bills at plan time; the
served container's argv shape is what the arms declare.
rad-convert's GPTQ path reads importance from an imatrix GGUF; the reference
container's --kld-record pass emits per-expert and per-hidden Grams as raw
recordings. This folds them into one v3 file: a single expert-major f32 entry
per MoE layer plus the hidden-width entries, every data offset a multiple of
general.alignment as the loader requires.
Refusals, each one a silent-worse outcome the swift15 campaign hit: an expert
with no gram of its own carries the layer's pooled diagonal and the count is
reported (a silent zero is a down-quantized expert nobody priced); expert grams
with no pooled gram, a gram short by one float (names the file), the same global
expert recorded twice with different numbers, an empty recording directory, and
an existing output. --min-rows sets the noise floor below which a recording is
dropped. rows is u64 so a corpus that outgrows u32 survives the header instead
of wrapping.
The swift15 rule (top 7% by imatrix energy, bar-only) left the plane size to the
distribution. The w5nl64 arm trades a *pinned host pool* against KL, so the
operator needs to say how many GiB of bf16 experts the pool can afford and have
the set land there: --target-count walks the ranked candidates until the byte
bill matches, keeps the bar-only pass for experts that are disproportionate for
their size, and reports the pool's ceiling instead of padding when the count
cannot be reached.
TP=2 cuts the expert plane into two 256-expert rank halves, and an odd count in
either half unbalances the placement, so the parity audit is a refusal (exit 2,
'PARITY LAW BROKEN'), not a silent fix-up: the driver stops rather than
converting a payload whose expert placement no one priced.
The w5nl64 campaign measures against the vendor checkpoint, not the Swift-1.5
finetune: the served payload (qwen3.8-next-flash-fp8-iq4r-moe) is quantized from
this base, so fidelity must be scored against the same reference. 335.3 GiB
across 10 shards.
Verifier gates on the index's own shard list, each shard's declared byte count,
and the Hub's declared total - a curl without -f writes the 401 body as the
'shard' and reports success. --fetch refuses when the destination cannot hold
the checkpoint, because a 335 GiB download that fills the disk mid-shard costs
the whole campaign window.
iptables -I prepends, so the install block's final -I of the blanket DROP
landed it at position 1, above all three ACCEPTs: every packet to
18093/tcp died there (live 2026-10-21: 159 packets on the DROP, 0 on each
ACCEPT). The host's own health probe timed out beside a LISTENing socket
and the whole dankclaw drain aborted at the alertmanager gate twice.
Install now rebuilds the tail: delete any blanket DROP wherever it sits,
keep the ACCEPTs at the head, append the DROP last - converges on hosts
poisoned by the old order. New bats case runs install_signal_relay for
real against an iptables sandbox seeded with the poisoned order (red on
the old code, green on this one).
Per Frank, frank-archon is public alongside dragoon/zealot: three-leg DB
trio (conn enable, model row ctx 200000, access_grant principal='*'),
verified as role=user. Token-mint traps recorded (cwd, PID-1 secret).
Closing the command-center security queue #70/#71/#72/#73/#74: rolling tags
let a rebuild move under a running service with nobody approving it, and
make a CVE audit ungradeable. 36 quadlet units across dankclaw, dankserver
and frankbook-pro now carry tag@sha256; localhost/* builds stay exempt
(their build is the pin). The pihole and copyparty pins pull the fixed
builds; valkey 9.1.2 (already running) and redis 7.4.11 are pinned in
place, vendor fixes TBD.
New gate tests/quadlet_image_pins.bats: no unpinned registry image, no
generator-rejected key anywhere. Three seams that pinned the old exact-tag
law (frankbook readarr, open_webui, seerr) now require tag AND digest -
strictly stronger. Quadlet Lifecycle docs: batch-abort law, entrypoint
flags-only law, pin law.
Three stacked lessons, each proved live 2026-10-21:
- the image's ENTRYPOINT is /bin/blackbox_exporter, so Exec= repeating the
binary made it a positional arg: 'unexpected /bin/blackbox_exporter'.
- the podman gateway 169.254.1.2 is NOT bindable from the host netns - it
exists only inside each rootless netns: 'bind: cannot assign requested
address'. The scrape path does not need it bound: a bridge container
dialing the gateway lands on the host LOOPBACK (proved 200 from the
prometheus container), so the exporter binds 127.0.0.1 + LAN.
- HealthCmd shell form ('/bin/sh -c "..."') is passed as one quoted token
and quadlet folds the following --health-interval/--health-retries (and
the image) into the command string; the plain exec form the sibling units
use is correct.
Prometheus.md bind law corrected (gateway is a forward-to-loopback, not a
bind address); deploy-remote gate stays loopback, which is now honest.
Third boot failure: the exporter exited 100 ms in with 'unknown long
flag --listen-address' (proved against the pinned image, plus --help).
The bind semantics were right; the flag name was invented. --web.listen-address
is repeatable - one per address - so the dual gateway+LAN bind stands.
Proven again pre-deploy with the real argv: the run starts and dies
only on the deliberate /dev/null placeholder config.
The deploy died at 'Unit blackbox-exporter.service not found': a
.container without Image= is not generated at all, and the header
argued for Network=host it never declared. Completes the unit (digest
pinned prom/blackbox-exporter, Network=host) and fixes the bind
topology the first cut got wrong:
- Prometheus is bridge-networked and the netavark hairpin gap bars it
from its own host's LAN IP, so a LAN-only prober bind is unscrapable
and a loopback-only bind was claimed to ride the gateway on faith.
The prober now binds 9115 on the podman gateway (scrape path) AND
the LAN (healthcheck/diagnostics); the listen_address line leaves
blackbox.yml so config and Exec cannot disagree.
- comfyui probe target http://comfyui:8188 -> http://127.0.0.1:8188:
bridge container names do not resolve in the host netns; the
published 0.0.0.0:8188 answers on loopback there.
- New seam: every observability unit must declare Image= (red on the
shipped file by name); copier test pinned module presence instead
of the removed listen line. Docs updated to the real topology.
PR #69 wired the copier, planner entry and health gate but missed the
three dispatch gates: the client valid_service, the remote
service_is_allowed, and deploy_service's case list. The client gate
rejected --service blackbox-exporter outright; even past it the
dispatcher's catch-all hard-fails on dankclaw (no compose since
2026-09-16). New seam pins the class: every dankclaw service the
remote gate allows must carry a deploy_service case - proven red by
planting a gated, undispatched name.
Measured 2026-10-21 with frank-dragoon live: MemAvailable 15.3 GiB,
iGPU busy 1.75 GiB. The Qwen-Image weight sets (13.3-14.2 GB) do not
fit that budget, and the page's 122 GB /system_stats figure is the
aperture, not headroom. Adds the fit table and the
stop-mask-render-verify procedure (the 3090 bench law, applied to the
APU); ComfyUI.md gotcha links it.
Prometheus/Observability/Provisioning pages: every job up after the
blackbox swap, the copier list learns copy_blackbox_config and the
alloy validate gate, the config table gains alloy + blackbox rows, and
the logs bullet states the stream-label law instead of the retired
'awaiting deploy' note.
- caddyfile seam learns the documented exception: URL-shaped targets
are legal only inside a prober job (state is per-job; the first cut
leaked it and passed a planted URL).
- observability seams: blackbox unit + planner entries, both copiers
executed for real in sandboxes, alloy engine-validation refusal
(image-gated), dashboard panel invariants.
- AGENTS.md: structural target downs retired.
copy_alloy_config now validates the rendered config with the very
alloy image that will run it, before installing: a misplaced attribute
crash-loops the unit behind a green deploy, and no textual seam can
see it. copy_blackbox_config ships the prober config; the planner maps
the new unit and shared/blackbox/** to the blackbox-exporter service,
ordered before prometheus.
Loki metrics queries group only by indexed stream labels, so the
journal-as-JSON stream made every per-container view structurally
empty (measured 6.5k lines/5min flowing while the board read zero).
A loki.process relabel stage now extracts CONTAINER_NAME into a
'container' label. alloy 1.21's json stage has no on_error attribute
(validated against the pinned image) — format_as_json guarantees the
line shape, so no error branch exists.
observability.json: stale 'structural downs' panel description fixed,
target census matches the post-blackbox job set. logs-overview.json:
panels grouped by the promoted container label instead of the absent
container_name, so the board stops reading empty beside live traffic.
Prometheus 3 refuses HTML scrape bodies, so the comfyui/services probe
jobs were structurally dead targets reading as outages. HTTP liveness
now rides a blackbox exporter (shared/blackbox/blackbox.yml, loopback
bind, http_2xx module); prometheus.yml probes SearXNG, OpenWebUI and
ComfyUI via __param_target relabeling. The unit is a boot-persistent,
self-healing root quadlet.
copy_grafana_provisioning named llm-fleet.json alone, so a new board in
shared/grafana/provisioning/dashboards/ silently never reached
/opt/grafana/provisioning. Measured 2026-10-21: the containers board was
merged, seam-tested, and documented as working since 2026-10-20, yet was
NEVER provisioned — its panel fixes were undeployable; the other four
boards existed on disk only because someone hand-copied them once (and
host-system.json there had gone stale against the repo). The copier now
globs the directory, and GRAFANA_PROV_ROOT lets the new bats case run
the copier for real into a sandbox (red against the old hand-listed
body, green against the glob). Vault: the copier law + the containers
incident on Observability Provisioning, the corrected dashboards claim
on Observability.
Note: a mktemp sandbox + trap RETURN races bats' trap machinery (the
dir vanishes mid-function); BATS_TEST_TMPDIR is the safe idiom.
Census: all 5 provisioned dashboards, 56 targets, every expression
re-evaluated against the live Prometheus/Loki with template vars expanded
(queries must run from the host: both APIs bind 127.0.0.1; from the dev box
that is connection-refused, which is NOT a dashboard defect).
- host-system 'Load 5m / core' was STRUCTURALLY EMPTY: bare '/' drops
non-matching pairs and the right side collapses to {job} while the left
carries {job,instance} -> 0 series forever. Fixed with / on(job); proven
live (2 series).
- containers 'Memory vs limit (%)' was STRUCTURALLY EMPTY: every
container_spec_memory_limit_bytes on this fleet is 0 (no cgroup limits),
so the '> 0' filter can never match. Rewritten to share-of-host-total via
label_replace join of the node exporter's MemTotal by host (the parens
form is a PromQL parse error: nothing may follow group_left); 62 series
live. Panel unit percentunit kept -> expression returns a fraction.
- llm-fleet 'Firing fleet alerts' is legitimately empty when healthy
(ALERTS carries only firing/pending series); description now says so,
per the Stale Signal rule that 0-series-on-ALERTS is not a defect.
logs-overview and observability: all targets already returning data.
Purpose-test pass over the audit findings (every verdict re-verified
against the file before touching it; three scout findings were hallucinated
— a nonexistent filename, my own changelog line, and sanctioned Eras rows):
- Observability Provisioning: destroyed the 63-line 'Current delivered
state' block — chain topology/relay/delivery restates [[Observability]]
and [[Alertmanager]] (those are the owners). Its three sole-custody laws
MOVED to Stale Signal Audit as 'Three rules that outlive the incident':
down-beside-healthy -> read lastError; Prometheus alert state in process
vs Alertmanager's --storage.path persistence; candidate-arm metrics are
never fleet state.
- RDNA4 MoE VRAM budget: kept the laws (registry over-prices the pool;
never reuse a B/token ratio across geometries; pin pool bytes
floor-sized), era-labelled the rates, destroyed the measured ladder
table for the superseded shape-G/bf16 rungs.
- Stale Signal Audit: new rules section + Provisioning Related link.
check-okf 109/0, vault links 0 unresolved, docs seams green.
The lane's child dlopens libxrt_core.so.2 from /opt/xilinx/lib/<multiarch>/
as well; our mounts filled only /opt/xilinx/xrt/lib and the install path.
Since the host's libxrt2 was reconfigured 2026-10-20 the child dies naming
the missing prefix ('No such library
/opt/xilinx/lib/x86_64-linux-gnu/libxrt_core.so.2') and the engine
supervisor takes the WHOLE container down on a dead lane — measured
2026-10-21: frank-dragoon boot-looped 88 times with chat dead while the
GPU lines looked healthy.
Each XRT library now mounts THREE times from the same real file (never a
symlink); the multiarch name derives from FRANK_DRAGOON_XRT_DIR's basename
so the bats fixture proves the derivation. Proven end-to-end before commit:
production-shaped scratch boot + the three extra mounts answers the full
boot ledger, npu_api buckets u0..u9, and /health ok at the 262144 ceiling.
A dead lane is not survivable on this engine — arm it only with all three
prefixes (page law updated).
Shape + harness contract on the endpoint page, AGENTS.md, and the OMP
Harness worked example recomputed to the live pair; the 400-zone law now
names the 85 %-comments vs 95 %-config disagreement as the recurring
cause and points at the seam that enforces the arithmetic.
The 200,704 ceiling squeezed the harness contextWindow down to 155,648
(the 400-zone fit law needs ~45k of gap under the ceiling), which is why
long dragoon sessions kept hitting the 400 zone: the harness clamps
max_tokens to window - est - 4096, its chars/4 est under-counts ~3 %,
and config.yml compacts at 95 % while the pair was designed for 85 %.
Ceiling -> checkpoint's native 262,144 (no RAM move: the KV pool stays
401,408 and is what the ledger bills; the band ctx <= pool <= slots*ctx
still holds). The harness window becomes 200,000 with a 62,144 gap.
The fit-law seam now reads the declared window from the vault page (the
one owner) and checks the arithmetic at the live 95 % trigger, so a
ceiling/window pair that drifts back together fails tests/run.sh
instead of a live session.
Two shipped-2026-10-20 bugs that left gpu-dankclaw, gpu-dankserver and
cadvisor-dankclaw DOWN beside a green deploy:
- gpu-exporter.py built one ThreadingHTTPServer per bind but ran
serve_forever() on servers[0] only: the LAN-class socket sat listen()ing
but never accept()ed, loopback deploy gate passed, every scrape that
reached the other bind timed out on the SYN backlog. Now one thread per
extra bind. Regression test spins --serve on two loopback binds and
requires BOTH to answer /metrics.
- cadvisor-dankclaw scraped its own host's LAN IP. Measured from inside the
Prometheus container: a bridge container cannot dial its own host's LAN
address at all (hairpin refuses published ports too); the podman gateway
answers. Target moved to host.containers.internal:8080; the seam that
pinned the LAN address (asserting the wrong doctrine) is re-pinned to the
gateway plus a negative pin against the own-host LAN dial. Observability
page records the measured reachability law replacing the inferred one.
deploy.sh's valid_service and deploy-remote.sh's service_is_allowed are
two gates for the same door, and they drifted: the remote gate gained
cadvisor (2026-10-20) and gpu-exporter (2026-10-21) while the client gate
never did, so `--service cadvisor|gpu-exporter` died with "is not deployed
on <target>" at the client — the first-ever deploy of a new service is
exactly the case that needs the client gate to know it. dankserver:prowlarr
(planner can emit it for dankserver) was remote-only too.
Adds the four missing pairs to the client gate and a deployment_flow
contract test that diffs the two extracted lists per target — the test the
cadvisor PR should have carried. searxng-network stays the one sanctioned
client-only exception (network unit, no remote handler), documented inline.
ls329 is the newest build on the linuxserver channel and Plex 1.43.5.11029
is their newest release; the live pin (ls325 / 1.43.4.10903) is two Plex
releases behind. Digest-verified at pin time: the tag's manifest digest
equals the registry's current 'latest' (sha256:8f655bad...e9e2334), the
same law the page records. Quadlet restart runs PMS's usual one-minor
schema migration on first boot; the deployer's /identity HTTPS gate
tolerates the first-boot migration up to 600 s.
Rollback: revert this commit (one line).
Add a host-process GPU exporter, scrape it on both servers, and graph it
on the host-system dashboard. The fleet has four GPUs across three vendors
of interest (2x R9700 on Dankserver, the 8060S iGPU + RTX 3090 on Dankclaw)
and nothing exposed them until now.
scripts/gpu-exporter.py reads each AMD card's amdgpu sysfs
(gpu_busy_percent / mem_info_vram_* / hwmon temp1_input / power1_average)
plus one nvidia-smi --query-gpu CSV, and emits unified gpu_* Prometheus
families (utilization_percent, memory_used/total_bytes, temperature_celsius,
power_watts) labelled gpu/vendor/model. Absent hardware yields ABSENT series,
never a fabricated 0: the R9700s expose no sysfs power, so they simply carry
no gpu_power_watts line. Both sources are injectable, so the seam runs on a
GPU-less dev box; live-verified on Dankserver's real sysfs.
It runs as a plain user unit (not a quadlet): it must read the host's own DRM
sysfs and shell to nvidia-smi, which a container cannot do without GPU
passthrough. Same posture as signal-relay. The base unit binds loopback; the
deploy drop-in appends the host LAN address, never a wildcard (matching
node-exporter/cadvisor). deploy_gpu_exporter installs/enables/health-gates it
on both hosts and is wired into the planner (both targets), the test-changed
router, and two Prometheus scrape jobs using the node/cadvisor reach rules
(gateway on Dankclaw, LAN on Dankserver).
Frank
dank