frank-primary: cutover state to Qwen3.8-Flash-Next Q8_0 (6 shards, PLE pinned, MTP4, sharp template)

- env: STATE_FILE -> r9v-state/q8xl; add dev-overlay adapter mount
  (quantized inject shard dequant, fork 363d606), FULL_MUTABLE_CACHE=0
  (80-slot tiered vs the overlay's uncensored 160-slot injection), and
  the qwen-sharp-v22.4.0.jinja chat template via the launcher's new
  R9V_CHAT_TEMPLATE_FILE/KWARGS passthrough (fork 30d4a32)
- serve/deploy: shard gates extended to SHARD4/5/6 (unset-safe, the
  uncensored rollback state carries none)
- seams: STATE_FILE/LLM_PRIMARY_STATE -> q8xl

scripts/deploy-remote.sh

@@ -1185,10 +1185,12 @@ deploy_llm_small() {
 
 # deploy_llm_primary — frank-primary is the R9V radiance-vLLM fork (docker-
 # pinned prebuilt, gfx1201, driven through the docker->podman shim) serving
-# Qwen3.8-Flash-Next-Uncensored IQ4_XS with the MTP4 drafter on the two
+# the unsloth Qwen3.8-Flash-Next Q8_0 payload (6 GGUF shards) with the MTP4
-# R9700s as a ROOT unit whose serve script execs the fork's launcher through
+# drafter on the two R9700s as a ROOT unit whose serve script execs the
-# rootless podman as dank. Same port and alias as the llama.cpp endpoint it
+# fork's launcher through rootless podman as dank. Same port and alias as
-# replaced (2026-09-27 cutover); the harness models.yml id is unchanged.
+# the endpoints it replaced (the 2026-09-27 cutover from uncensored
+# IQ4_XS); the harness models.yml id is unchanged. The uncensored state dir
+# (/home/dank/r9v-state/uncensored) stays on disk as the rollback target.
 #
 # Gates: the on-host R9V setup state must exist and be ready (it carries the
 # pinned image digest and every payload path), the pinned runtime image must
@@ -1196,7 +1198,7 @@ deploy_llm_small() {
 # PLE payload must match its pinned byte count, and every GGUF shard + the
 # mmproj must exist (per-file sha256 verification was setup's job at fetch
 # time, never re-hashed here).
-LLM_PRIMARY_STATE=/home/dank/r9v-state/uncensored/setup.json
+LLM_PRIMARY_STATE=/home/dank/r9v-state/q8xl/setup.json
 LLM_PRIMARY_CONTAINER=frank-primary-r9v
 
 deploy_llm_primary() {
@@ -1216,7 +1218,9 @@ config = state.get("config") or {}
 if state.get("ready") is not True:
     sys.exit("R9V setup state is not ready")
 for key in ("R9V_IMAGE", "R9V_MODEL_DIR", "R9V_PLE_PATH", "R9V_PLE_EXPECTED_BYTES",
-            "R9V_TARGET_REL", "R9V_TARGET_SHARD2_REL", "R9V_TARGET_SHARD3_REL", "R9V_MMPROJ_REL"):
+            "R9V_TARGET_REL", "R9V_TARGET_SHARD2_REL", "R9V_TARGET_SHARD3_REL",
+            "R9V_TARGET_SHARD4_REL", "R9V_TARGET_SHARD5_REL", "R9V_TARGET_SHARD6_REL",
+            "R9V_MMPROJ_REL"):
     print(f"{key}={shlex.quote(config.get(key, ''))}")
 PY
 )" || return 1
@@ -1239,7 +1243,9 @@ PY
         fi
     fi
     local rel
-    for rel in "$R9V_TARGET_REL" "$R9V_TARGET_SHARD2_REL" "$R9V_TARGET_SHARD3_REL" "$R9V_MMPROJ_REL"; do
+    for rel in "$R9V_TARGET_REL" "${R9V_TARGET_SHARD2_REL:-}" "${R9V_TARGET_SHARD3_REL:-}" \
+        "${R9V_TARGET_SHARD4_REL:-}" "${R9V_TARGET_SHARD5_REL:-}" "${R9V_TARGET_SHARD6_REL:-}" \
+        "$R9V_MMPROJ_REL"; do
         [[ -z "$rel" ]] && continue
         if [[ ! -f "$R9V_MODEL_DIR/$rel" ]]; then
             echo "ERROR: payload file missing: $R9V_MODEL_DIR/$rel" >&2

scripts/frank-primary.env

@@ -1,15 +1,16 @@
 # frank-primary.env — R9V radiance-vLLM production overrides (dankserver :7770).
 #
-# The full tuned engine shape (104 R9V_* keys: tiered-expert placement,
+# The full tuned engine shape (tiered-expert placement, CPU offload budgets
-# CPU offload budgets 146.25/78.75 GiB per rank, CED on, MTP4, TP=2,
+# per rank, MTP4 drafter, TP=2, qwen3_coder tool parser, 131072 model len,
-# qwen3_coder tool parser, 196608 model len) lives on-host in the R9V setup
+# 80-slot tiered expert cache, PLE pinned residency 54.4 GB) lives on-host
-# state — /home/dank/r9v-state/uncensored/setup.json — written by the
+# in the R9V setup state — /home/dank/r9v-state/q8xl/setup.json — written by
-# profile's setup run (PASS=28/FAIL=0, all 20 artifacts sha256-verified).
+# the q8 profile's setup run (ready: true; doctor PASS, artifacts
-# serve-frank-primary.sh layers that state first and THIS FILE second, so the
+# sha256-verified at fetch time). serve-frank-primary.sh layers that state
-# keys below win. Never copy tuned values from the state into this file —
+# first and THIS FILE second, so the keys below win. Never copy tuned values
-# the state is the single source of truth for engine shape.
+# from the state into this file — the state is the single source of truth
+# for engine shape.
 R9V_HOME=/home/dank/GitPackages/radiance-vllm-qwen4exp
-R9V_STATE_FILE=/home/dank/r9v-state/uncensored/setup.json
+R9V_STATE_FILE=/home/dank/r9v-state/q8xl/setup.json
 # Fleet identity: LAN bind, the single frank-primary port, the harness's
 # model id, our own container name (the launcher default r9v-qwen38-flash-next
 # is the scratch identity).
@@ -18,3 +19,23 @@ R9V_HOST_PORT=7770
 R9V_SERVED_MODEL_NAME=frank-primary
 R9V_CONTAINER_NAME=frank-primary-r9v
 R9V_EXPECTED_GPU_BDFS=0000:03:00.0,0000:07:00.0
+# The Q8_0 payload quantizes BOTH fused hyper-connection shards (the IQ4_XS
+# package ships inject as F32), so the shard-uniformity check rejects the
+# fused projection at load. This dev-overlay adapter dequantizes a quantized
+# inject shard in-stream (fork commit 363d606): opt-in mount, uncensored
+# payloads pass through unchanged.
+R9V_DEV_QWEN4_EXP_ADAPTER_PY=/home/dank/GitPackages/radiance-vllm-qwen4exp/runtimes/dev-overrides/qwen4_exp.py
+# The runtime overlay's env.always injects R9V_FULL_MUTABLE_CACHE=1 — the
+# qualified 160-slot uncensored IQ4_XS cache geometry — which the q8
+# profile's 80-slot tiered cache fails with a baseline mismatch. This flips
+# that injection off for this payload (launch.sh passes it after the overlay
+# envs; docker last-wins).
+R9V_FULL_MUTABLE_CACHE=0
+# The Qwen-Sharp template (the frank-large fleet law): folds reasoning
+# efforts (high/xhigh/max -> xhigh, minimal/low -> low) and defaults
+# preserve_thinking true. --chat-template wins over the tokenizer dir's
+# metadata/chat_template.jinja; request-level kwargs still win per request.
+# Braces in values get the same single-quote law as parentheses (source
+# parses assignment values).
+R9V_CHAT_TEMPLATE_FILE=/home/dank/serverAppConfigs/scripts/qwen-sharp-v22.4.0.jinja
+R9V_CHAT_TEMPLATE_KWARGS='{"preserve_thinking":true}'

scripts/serve-frank-primary.sh

@@ -1,7 +1,8 @@
 #!/usr/bin/env bash
 # serve-frank-primary.sh — frank-primary endpoint (R9V radiance-vLLM fork,
-# Qwen3.8-Flash-Next-Uncensored IQ4_XS + MTP4 drafter, 2x R9700, dankserver
+# unsloth Qwen3.8-Flash-Next Q8_0 + MTP4 drafter, 2x R9700, dankserver
-# :7770).
+# :7770; 2026-09-27 cutover from the uncensored IQ4_XS payload, whose state
+# dir stays on disk as the rollback target).
 #
 # The engine is the upstream R9V launcher (scripts/launch.sh in the fork
 # checkout), NOT a hand-frozen podman argv: launch.sh owns the resolved
@@ -20,10 +21,10 @@
 #
 # The root unit (scripts/frank-primary.service, User=dank,
 # LimitMEMLOCK=infinity) grants the pin budget for the tiered-expert UVA
-# host copy (~43 GiB pinned) and the 16 GiB PLE pinned reserve. A plain
+# host copy and the PLE pinned residency (54.4 GB on the q8 payload). A
-# shell session also works (proven 2026-09-27: podman rootless inherits
+# plain shell session also works (proven 2026-09-27: podman rootless
-# usable limits here), but the unit is the fleet-parity boot-persistent
+# inherits usable limits here), but the unit is the fleet-parity
-# shape.
+# boot-persistent shape.
 set -euo pipefail
 
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
@@ -66,7 +67,7 @@ export PATH="$SHIM_DIR:$PATH"
 STATE_FILE="$R9V_STATE_FILE"
 if [ ! -f "$STATE_FILE" ]; then
     echo "serve-frank-primary: R9V setup state missing: $STATE_FILE" >&2
-    echo "  run the profile setup first (r9v setup qwen38-mtp4-uncensored ...)" >&2
+    echo "  run the profile setup first (tools/setup_profile.py setup --state-dir ...)" >&2
     exit 1
 fi
 STATE_EXPORTS="$(python3 - "$STATE_FILE" <<'PY'
@@ -92,12 +93,13 @@ eval "$STATE_EXPORTS"
 # digest but preserves the manifest digest in RepoDigests — the shim's ID law).
 if ! podman images --format '{{.RepoDigests}}' | grep -qF "${R9V_IMAGE#sha256:}"; then
     echo "serve-frank-primary: pinned R9V image absent: $R9V_IMAGE" >&2
-    echo "  load the bundle: ./r9v setup qwen38-mtp4-uncensored --state-dir ..." >&2
+    echo "  load the pinned image bundle into podman rootless storage" >&2
     exit 1
 fi
 
 # PLE payload at its pinned byte count (stat -c %s, never wc -c), plus the
-# three target GGUF shards and the mmproj the launcher mounts.
+# six target GGUF shards (mradermacher-style split parts) and the mmproj
+# the launcher mounts.
 PLE_BYTES="${R9V_PLE_EXPECTED_BYTES:-}"
 if [ -n "$PLE_BYTES" ]; then
     ACTUAL="$(stat -c %s "$R9V_PLE_PATH" 2>/dev/null || echo 0)"
@@ -112,7 +114,9 @@ else
         exit 1
     fi
 fi
-for rel in "$R9V_TARGET_REL" "$R9V_TARGET_SHARD2_REL" "$R9V_TARGET_SHARD3_REL" "$R9V_MMPROJ_REL"; do
+for rel in "$R9V_TARGET_REL" "${R9V_TARGET_SHARD2_REL:-}" "${R9V_TARGET_SHARD3_REL:-}" \
+    "${R9V_TARGET_SHARD4_REL:-}" "${R9V_TARGET_SHARD5_REL:-}" "${R9V_TARGET_SHARD6_REL:-}" \
+    "$R9V_MMPROJ_REL"; do
     if [ -z "$rel" ]; then continue; fi
     if [ ! -f "$R9V_MODEL_DIR/$rel" ]; then
         echo "serve-frank-primary: payload file missing: $R9V_MODEL_DIR/$rel" >&2

tests/frank_primary_unit.bats

@@ -65,11 +65,11 @@ REMOTE_DEPLOY="$REPO_ROOT/scripts/deploy-remote.sh"
 }
 
 @test "the tuned engine shape comes from the on-host setup state, not the repo" {
-    # The profile setup qualified a 104-key shape (tiered-expert placement,
+    # The profile setup qualified the tuned shape (tiered-expert placement,
-    # per-rank offload budgets, CED, MTP4, parser wiring). Copying it into
+    # per-rank offload budgets, MTP4, parser wiring). Copying it into
     # the repo would drift; the serve script layers it and skips keys the
     # env file defines, so production overrides always win.
-    grep -q 'R9V_STATE_FILE=/home/dank/r9v-state/uncensored/setup.json' "$ENV_FILE"
+    grep -q 'R9V_STATE_FILE=/home/dank/r9v-state/q8xl/setup.json' "$ENV_FILE"
     grep -q 'state must carry R9V_MODEL_DIR' "$SERVE"
     grep -q 'is not ready' "$SERVE"
     grep -q 'key not in env_keys' "$SERVE"
@@ -122,7 +122,7 @@ REMOTE_DEPLOY="$REPO_ROOT/scripts/deploy-remote.sh"
     # the PLE payload is off its pinned byte count, or when any GGUF shard
     # or the mmproj is missing — then clears stale containers, waits out
     # the port, relinks/enables/restarts the root unit and health-gates it.
-    grep -q 'LLM_PRIMARY_STATE=/home/dank/r9v-state/uncensored/setup.json' "$REMOTE_DEPLOY"
+    grep -q 'LLM_PRIMARY_STATE=/home/dank/r9v-state/q8xl/setup.json' "$REMOTE_DEPLOY"
     grep -q 'LLM_PRIMARY_CONTAINER=frank-primary-r9v' "$REMOTE_DEPLOY"
     grep -q 'is not ready' "$REMOTE_DEPLOY"
     grep -q "podman images --format '{{.RepoDigests}}'" "$REMOTE_DEPLOY"