SyslogSolution/prose-contracts
7
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

fix: align contracts to 4-name LiteLLM registry (2026-09-12)
PR Pipeline — Authorize → Validate → Review → Merge / auth (push) Successful in 3s
Details
PR Pipeline — Authorize → Validate → Review → Merge / validate (push) Successful in 3s
Details
PR Pipeline — Authorize → Validate → Review → Merge / lint (push) Successful in 2s
Details
PR Pipeline — Authorize → Validate → Review → Merge / ai-review (push) Successful in 2s
Details
PR Pipeline — Authorize → Validate → Review → Merge / gate (push) Successful in 1s
Details

Browse Source 
- Remove retired names (qwen3.6-27B-code, qwen3.6-35B-udq4) from live alias claims
- Update Strix Halo model to Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (strix-moe, 256K ctx)
- Fix litellm-health step 7 probe to gpu-vision (monitor key scoped)
- Move qwen3.6-27B-code/35B-udq4 from raw-but-live to non-resolving in audit
- Fold in pm2-self-heal: remove spoton-service (live PM2 set is 4/4)
- Update hermes templates, key enforcement, timeout tables to live names
This commit is contained in:
root
2026-09-12 21:39:54 +00:00
parent b3644f0292
commit 8210fd905c
11 changed files with 27 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gpu-fleet.prose.md
+5 -5
View File
@@ -91,7 +91,7 @@ Single source of truth for models, aliases, rpm caps, weights and fallback chain
CT 116 `/opt/inference-harness/litellm_config.yaml`. Do not duplicate those values in
contracts — read them there.
**Backward compatibility**: Old model-specific names (qwen3.6-27B-code, qwen3.5-9b-it) still work
**No backward compatibility**: Old model-specific names (qwen3.6-27B-code, qwen3.5-9b-it) are retired as of 2026-09-12
but are deprecated for agent configs. Only the stable aliases survive model swaps. `gemma-4-12b`
is retired and returns 400 `Invalid model name`.
@@ -207,7 +207,7 @@ If no SSH access, send Zulip DM via abiba-bot with vault update instructions.
- **Router startup race**: Compose router.py doesn't call load_roster(). Reload thread sleeps 30s first.
Fix: trigger roster reload via SSH after restart, or rebuild image with startup load_roster().
- **LiteLLM /metrics**: Requires auth. Prometheus uses `/health/liveliness` as workaround.
- **Strix Halo GPU**: Vulkan is the working backend (ROCm/HIP path abandoned — HSA runtime blocked on Debian 13). Build at `/root/llama.cpp/build-vk/`, commit `4fc4ec5` (2026-07-01), ggml 0.15.3 shared-lib arch. Mesa RADV 25.0.7, KHR_coopmat fast path active. ~70 tok/s gen, 532 tok/s prompt. Service: `strix-server.service` on port 8080, model: `qwen3.6-35B-udq4`, alias `strix-moe`, 128K context, flash-attn + q4 KV, multimodal (mmproj loaded).
- **Strix Halo GPU**: Vulkan is the working backend (ROCm/HIP path abandoned — HSA runtime blocked on Debian 13). Build at `/root/llama.cpp/build-vk/`, commit `4fc4ec5` (2026-07-01), ggml 0.15.3 shared-lib arch. Mesa RADV 25.0.7, KHR_coopmat fast path active. ~70 tok/s gen, 532 tok/s prompt. Service: `strix-server.service` on port 8080, model: `Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf`, alias `strix-moe`, 256K context (n_ctx 262144), --parallel 2 --kv-unified, flash-attn + q4 KV, multimodal (mmproj loaded).
- **Port conflict detection (2026-07-05)**: All 3 GPU wrappers now detect ghost processes squatting port 8080 before starting. `.8` and `.110` use inline pre-start check in `llama-wrapper.sh`; `.15` uses `/usr/local/bin/port-cleanup.sh` ExecStartPre. Replaces the blanket `pkill -9 -x llama-server` on .15 which would kill ALL llama-server instances regardless of port. Ghost detection was the root cause of .8 crash-looping for 27+ restarts (stale pid 25836 squatting 8080 after OOM kill).
- **Strix Halo thermal safeguard (2026-07-02)**: `strix-server.service` has `-n 8192` (hard generation cap per request). Without it, `--predict` defaults to -1 (infinity) — a runaway request from .123 (old Mumuni CT114 — now inside Abiba CT100 at .24) decoded 39,868 tokens over 24 min, pushing Tctl to 98°C (crit 89.8°C) and throttling 70→29 t/s. The cap bounds worst-case generation to ~5 min. Do NOT remove `-n` without a replacement ceiling. Sustained load hits ~84°C even at 92s; the APU is fanless/low-flow. Clients MUST also set `max_tokens`.
- **Port 8080 firewall**: amdpve iptables restricts 8080 to 192.168.68.116 (LiteLLM/router host) only. All inbound connections are from .116 (LiteLLM proxied via nginx). Localhost curls hang (SYN dropped). Always test from .116.
@@ -223,10 +223,10 @@ If no SSH access, send Zulip DM via abiba-bot with vault update instructions.
| GPU | Model | Gen tok/s | Prompt tok/s | Baseline | Context |
|-----|-------|-----------|--------------|----------|---------|
| RTX 3090 (.8) | Qwen3.8-27B-Uncensored-Q4_K_M | **TBD** | — | — | **128K** |
| Strix Halo (.15) | qwen3.6-35B-udq4 | **65** | 140 | — | **128K** |
| Strix Halo (.15) | Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (strix-moe) | **65** | 140 | — | **256K** |
Benchmarks from 2026-07-17. Strix Halo model: qwen3.6-35B-udq4. RTX 5070 MTP provides 2.7x speedup over pre-upgrade 70 tok/s.
All 3 GPUs now at 128K context (2026-07-17, reduced from 256K for stability).
Benchmarks from 2026-07-17. Strix Halo model: Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf (alias strix-moe), n_ctx 262144, --parallel 2 --kv-unified. RTX 5070 MTP provides 2.7x speedup over pre-upgrade 70 tok/s.
GPU contexts: RTX 3090 (.8) and RTX 5070 (.110) at 128K; Strix Halo (.15) at 256K (2026-09-12).
Benchmarks run through LiteLLM proxy (192.168.68.116:4000) every 5 minutes.
Degradation alerts fire at 30% (warning) and 50% (critical) below baseline.