Compare commits
4 Commits
feat/nix-d
...
feat/herme
| Author | SHA1 | Date | |
|---|---|---|---|
| bb64234223 | |||
| 9fbfc1dc51 | |||
| 7e3afe6630 | |||
| 30f8ca3863 |
203
assets/ai-optimizer/CRON_EXECUTION_PROMPT.md
Normal file
203
assets/ai-optimizer/CRON_EXECUTION_PROMPT.md
Normal file
@@ -0,0 +1,203 @@
|
||||
# AI Model Optimization Cron Job - EXECUTION PROMPT
|
||||
|
||||
**When this cron runs, follow these instructions exactly:**
|
||||
|
||||
---
|
||||
|
||||
## Your Role
|
||||
|
||||
You are an AI model optimization agent. Your task is to find the best ollama/llama.cpp configuration for maximum context size and hardware utilization.
|
||||
|
||||
**Hardware:**
|
||||
- 2× AMD MI50 GPUs (32GB VRAM each, 64GB total)
|
||||
- 128GB system RAM
|
||||
- ROCm: HSA_OVERRIDE_GFX_VERSION=9.0.6, HIP_VISIBLE_DEVICES=0,1
|
||||
|
||||
---
|
||||
|
||||
## File Locations
|
||||
|
||||
```
|
||||
STATE: /opt/data/infra/assets/ai-optimizer/state.json
|
||||
RESULTS: /opt/data/infra/assets/ai-optimizer/results.csv
|
||||
INFRA_REPO: /opt/data/infra
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Model Queues
|
||||
|
||||
### GPU Track (Coding - prioritize speed + context on GPU)
|
||||
1. `devstral-small-2:24b`
|
||||
2. `qwen2.5-coder:32b`
|
||||
3. `codellama:34b-instruct`
|
||||
|
||||
### RAM Track (Knowledge - prioritize max context)
|
||||
1. `qwen2.5:72b`
|
||||
2. `nemotron-3-nano:30b`
|
||||
3. `mixtral:8x7b-instruct`
|
||||
|
||||
---
|
||||
|
||||
## Context Steps (in order)
|
||||
```
|
||||
[32768, 65536, 98304, 131072, 163840, 200704, 262144, 327680]
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Each Run - Step by Step
|
||||
|
||||
### 1. Read State
|
||||
```bash
|
||||
cd /opt/data/infra
|
||||
cat assets/ai-optimizer/state.json
|
||||
```
|
||||
|
||||
### 2. Determine Next Test
|
||||
- Read `track` (gpu or ram)
|
||||
- Read `current_model` from queue at `model_index`
|
||||
- Read `current_config` for parameters to test
|
||||
- Select next context step from `context_steps` based on `phase`
|
||||
|
||||
### 3. Pull Model (if needed)
|
||||
```bash
|
||||
docker exec ollama ollama list | grep -q "<model>" || docker exec ollama ollama pull <model>
|
||||
```
|
||||
|
||||
### 4. Create Test Modelfile
|
||||
```bash
|
||||
docker exec ollama bash -c "cat <<EOF > /root/.ollama/test_${model}.modelfile
|
||||
FROM ${model}
|
||||
PARAMETER num_ctx ${current_config.num_ctx}
|
||||
PARAMETER num_gpu ${current_config.num_gpu}
|
||||
PARAMETER flash_attn ${current_config.flash_attn}
|
||||
PARAMETER num_predict 4096
|
||||
PARAMETER num_keep 1024
|
||||
PARAMETER repeat_penalty 1.1
|
||||
EOF"
|
||||
|
||||
docker exec ollama ollama create test-model -f /root/.ollama/test_${model}.modelfile
|
||||
```
|
||||
|
||||
### 5. Run Benchmark
|
||||
```bash
|
||||
# Warm up
|
||||
docker exec ollama ollama run test-model "Hello" > /dev/null
|
||||
|
||||
# Coding prompt
|
||||
START=$(date +%s%N)
|
||||
docker exec ollama ollama run test-model "Write a Python async context manager that retries a function with exponential backoff, max 5 retries, and logs each attempt using structlog. Include type hints."
|
||||
END=$(date +%s%N)
|
||||
|
||||
# Calculate tokens/sec from output
|
||||
```
|
||||
|
||||
### 6. Measure VRAM (if possible)
|
||||
```bash
|
||||
# Try host first
|
||||
rocm-smi --showmeminfo vram 2>/dev/null || \
|
||||
# Try via docker
|
||||
docker exec --privileged ollama rocm-smi --showmeminfo vram 2>/dev/null || \
|
||||
# Fallback
|
||||
echo "VRAM measurement unavailable"
|
||||
```
|
||||
|
||||
### 7. Record Results
|
||||
- Parse tokens/sec from ollama output
|
||||
- Record VRAM/RAM usage
|
||||
- Determine if this is best config so far for this model
|
||||
- Update `best_configs` if tokens/sec improved or context increased
|
||||
|
||||
### 8. Update State
|
||||
```python
|
||||
# Logic:
|
||||
if test_successful:
|
||||
if context_step < max_reached:
|
||||
phase = "context_scaling"
|
||||
current_config.num_ctx = next_context_step
|
||||
else:
|
||||
# Move to next model
|
||||
model_index += 1
|
||||
phase = "context_scaling"
|
||||
current_config.num_ctx = context_steps[0]
|
||||
else:
|
||||
# OOM or error - record last good as best
|
||||
best_configs[track][current_model] = last_good_config
|
||||
model_index += 1
|
||||
phase = "context_scaling"
|
||||
```
|
||||
|
||||
### 9. Commit to Repo
|
||||
```bash
|
||||
cd /opt/data/infra
|
||||
git add assets/ai-optimizer/
|
||||
git commit -m "ai-optimizer: tested ${model} at ${num_ctx} ctx - ${status}"
|
||||
git push origin master
|
||||
```
|
||||
|
||||
### 10. Matrix Notification (if available)
|
||||
```python
|
||||
import os
|
||||
if os.getenv("MATRIX_HOME_SERVER") and os.getenv("MATRIX_ACCESS_TOKEN"):
|
||||
# Send notification to Matrix room
|
||||
# Room ID from env or config
|
||||
pass
|
||||
# Else: silent
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Stop Conditions
|
||||
|
||||
1. All models in both queues have `best_configs` recorded
|
||||
2. Manual intervention needed (error in state.json `error` field)
|
||||
3. No progress for 3 consecutive runs (stuck)
|
||||
|
||||
---
|
||||
|
||||
## Error Handling
|
||||
|
||||
If any step fails:
|
||||
1. Log error to state.json: `"error": {"message": "...", "timestamp": "..."}`
|
||||
2. Do NOT increment model_index (retry next run)
|
||||
3. Commit state with error field
|
||||
4. Exit gracefully
|
||||
|
||||
---
|
||||
|
||||
## Important Notes
|
||||
|
||||
- **No num_parallel**: Do not use this parameter
|
||||
- **Two tracks**: Complete GPU track first, then RAM track
|
||||
- **Backend**: Start with ollama, llama.cpp testing is optional (requires uncommenting in compose.yml)
|
||||
- **Host access**: Some commands need host - use docker exec or SSH if available
|
||||
- **Ask before deploy**: If config changes needed in NixOS modules, show diff and wait for user confirmation before `nh os switch`
|
||||
|
||||
---
|
||||
|
||||
## Example State Transitions
|
||||
|
||||
**Start:**
|
||||
```json
|
||||
{"track": "gpu", "model_index": 0, "current_model": "devstral-small-2:24b", "current_config": {"num_ctx": 32768, ...}}
|
||||
```
|
||||
|
||||
**After successful test at 32k:**
|
||||
```json
|
||||
{"track": "gpu", "model_index": 0, "current_model": "devstral-small-2:24b", "current_config": {"num_ctx": 65536, ...}}
|
||||
```
|
||||
|
||||
**After OOM at 131k:**
|
||||
```json
|
||||
{
|
||||
"track": "gpu",
|
||||
"model_index": 1,
|
||||
"current_model": "qwen2.5-coder:32b",
|
||||
"best_configs": {
|
||||
"gpu": {
|
||||
"devstral-small-2:24b": {"num_ctx": 98304, "num_gpu": 99, "tokens_per_sec": 11.2}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
283
assets/ai-optimizer/CRON_JOB_DRAFT.md
Normal file
283
assets/ai-optimizer/CRON_JOB_DRAFT.md
Normal file
@@ -0,0 +1,283 @@
|
||||
# AI Model Optimization Cron Job
|
||||
|
||||
**Goal:** Find optimal configurations for maximum context size with full hardware utilization.
|
||||
|
||||
**Hardware:**
|
||||
- 2× AMD MI50 GPUs (32GB VRAM each, 64GB total)
|
||||
- 128GB system RAM
|
||||
- ROCm: HSA_OVERRIDE_GFX_VERSION=9.0.6, HIP_VISIBLE_DEVICES=0,1
|
||||
|
||||
---
|
||||
|
||||
## Model Queue
|
||||
|
||||
### GPU-Optimized (Coding - prioritize speed + context on GPU)
|
||||
1. `devstral-small-2:24b` - Best coding model
|
||||
2. `qwen2.5-coder:32b` - Strong coder, fits on GPU+offload
|
||||
3. `codellama:34b-instruct` - Legacy but solid
|
||||
|
||||
### RAM-Optimized (Knowledge - prioritize max context, accept slower)
|
||||
1. `qwen2.5:72b` - Best knowledge, needs heavy offload
|
||||
2. `nemotron-3-nano:30b` - Good general knowledge
|
||||
3. `mixtral:8x7b-instruct` - MoE, efficient for knowledge
|
||||
|
||||
---
|
||||
|
||||
## Optimization Strategy
|
||||
|
||||
**Two separate tracks:**
|
||||
|
||||
### Track A: GPU-Focused (Coding)
|
||||
```
|
||||
Baseline: num_ctx=32768, num_gpu=99, flash_attn=true
|
||||
Steps:
|
||||
1. Increase context: 32k → 65k → 98k → 131k → 163k
|
||||
2. At each step, verify VRAM usage < 60GB (leave headroom)
|
||||
3. If OOM: reduce num_gpu until stable, record best
|
||||
4. Measure tokens/sec - if < 5 tok/s, consider context too high
|
||||
```
|
||||
|
||||
### Track B: RAM-Focused (Knowledge)
|
||||
```
|
||||
Baseline: num_ctx=65536, num_gpu=50, flash_attn=true
|
||||
Steps:
|
||||
1. Increase context: 65k → 131k → 200k → 262k → 327k
|
||||
2. Allow heavy RAM offload (system RAM up to 100GB)
|
||||
3. If OOM: reduce context or num_gpu
|
||||
4. Speed less critical - focus on max stable context
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Backend-Specific Configs
|
||||
|
||||
### Ollama (Modelfile parameters)
|
||||
```
|
||||
PARAMETER num_ctx <value>
|
||||
PARAMETER num_gpu <layers>
|
||||
PARAMETER flash_attn true/false
|
||||
PARAMETER num_predict 4096
|
||||
PARAMETER num_keep 1024
|
||||
PARAMETER repeat_penalty 1.1
|
||||
```
|
||||
|
||||
### Llama.cpp (CLI flags)
|
||||
```
|
||||
--ctx-size <value>
|
||||
--n-gpu-layers <layers>
|
||||
--flash-attn on/off
|
||||
--n-predict 4096
|
||||
--batch-size 4096
|
||||
--ubatch-size 512
|
||||
--cache-type-k f16
|
||||
--cache-type-v f16
|
||||
--split-mode layer
|
||||
--no-mmap
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Host Test Instructions
|
||||
|
||||
**The cron runs inside the hermes container. Some tests require host access:**
|
||||
|
||||
### 1. VRAM Monitoring (HOST)
|
||||
```bash
|
||||
# Run on host to check VRAM usage during/after benchmark
|
||||
sudo rocm-smi --showmeminfo vram
|
||||
|
||||
# Or via docker exec if rocm-smi available in container
|
||||
docker exec --privileged ollama rocm-smi --showmeminfo vram
|
||||
```
|
||||
|
||||
### 2. Running Ollama Benchmarks (CONTAINER)
|
||||
```bash
|
||||
# Pull model
|
||||
docker exec ollama ollama pull <model>
|
||||
|
||||
# Create custom modelfile
|
||||
docker exec ollama bash -c 'cat <<EOF > /root/.ollama/test.modelfile
|
||||
FROM <model>
|
||||
PARAMETER num_ctx 65536
|
||||
PARAMETER num_gpu 99
|
||||
PARAMETER flash_attn true
|
||||
EOF'
|
||||
|
||||
# Create model from modelfile
|
||||
docker exec ollama ollama create test-model -f /root/.ollama/test.modelfile
|
||||
|
||||
# Run benchmark (warm model first)
|
||||
docker exec ollama ollama run test-model "Write a Python async context manager with exponential backoff"
|
||||
|
||||
# Cleanup
|
||||
docker exec ollama ollama rm test-model
|
||||
```
|
||||
|
||||
### 3. Running Llama.cpp Benchmarks (CONTAINER - needs llama.cpp container)
|
||||
```bash
|
||||
# Uncomment llama_cpp_devstral in compose.yml first
|
||||
# Then rebuild: sudo nh os switch --flake .#lazyworkhorse
|
||||
|
||||
# Test via HTTP API
|
||||
curl http://localhost:8300/v1/completions \
|
||||
-H "Content-Type: application/json" \
|
||||
-d '{
|
||||
"model": "devstral-2-small-llama_cpp",
|
||||
"prompt": "Write a Python function",
|
||||
"max_tokens": 100
|
||||
}'
|
||||
```
|
||||
|
||||
### 4. Deploying Changes (HOST via ai-worker)
|
||||
```bash
|
||||
# After optimization, commit results
|
||||
cd /home/ai-worker/infra
|
||||
git add assets/ai-optimizer/
|
||||
git commit -m "ai-optimizer: new best config for <model>"
|
||||
git push
|
||||
|
||||
# If config changes needed in ollama_init_custom_models.nix:
|
||||
# 1. Edit the file
|
||||
# 2. nixpkgs-fmt .
|
||||
# 3. Show diff to user
|
||||
# 4. Wait for confirmation
|
||||
# 5. sudo nh os switch --flake .#lazyworkhorse
|
||||
```
|
||||
|
||||
### 5. Accessing Host from Hermes Container
|
||||
```bash
|
||||
# SSH to host as ai-worker (key should be mounted)
|
||||
ssh -i /path/to/key ai-worker@host.docker.internal
|
||||
|
||||
# Or via docker socket if mounted
|
||||
# (not recommended for security)
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Benchmark Prompts
|
||||
|
||||
### Coding (Track A)
|
||||
```
|
||||
"Write a Python async context manager that retries a function with exponential backoff, max 5 retries, and logs each attempt using structlog. Include type hints and error handling."
|
||||
```
|
||||
|
||||
### Knowledge (Track B)
|
||||
```
|
||||
"Explain the complete memory hierarchy in modern GPUs, from registers through L1/L2 caches to VRAM, and how data moves between them during matrix multiplication. Include bandwidth considerations for each level."
|
||||
```
|
||||
|
||||
### Measurement
|
||||
- Tokens per second (generation speed)
|
||||
- Time to first token (latency)
|
||||
- VRAM usage (via rocm-smi)
|
||||
- System RAM usage (via free -h)
|
||||
- Context success (did it complete without OOM?)
|
||||
|
||||
---
|
||||
|
||||
## State File Structure
|
||||
|
||||
`/opt/data/infra/assets/ai-optimizer/state.json`
|
||||
|
||||
```json
|
||||
{
|
||||
"track": "gpu",
|
||||
"current_model": "devstral-small-2:24b",
|
||||
"model_index": 0,
|
||||
"phase": "context_scaling",
|
||||
"backend": "ollama",
|
||||
"current_config": {
|
||||
"num_ctx": 65536,
|
||||
"num_gpu": 99,
|
||||
"flash_attn": true
|
||||
},
|
||||
"best_configs": {
|
||||
"gpu": {
|
||||
"devstral-small-2:24b": {
|
||||
"backend": "ollama",
|
||||
"num_ctx": 131072,
|
||||
"num_gpu": 99,
|
||||
"flash_attn": true,
|
||||
"tokens_per_sec": 12.5,
|
||||
"vram_used_gb": 58.2,
|
||||
"tested_at": "2026-04-28T17:00:00Z"
|
||||
}
|
||||
},
|
||||
"ram": {}
|
||||
},
|
||||
"completed_models": [],
|
||||
"gpu_queue": ["devstral-small-2:24b", "qwen2.5-coder:32b", "codellama:34b-instruct"],
|
||||
"ram_queue": ["qwen2.5:72b", "nemotron-3-nano:30b", "mixtral:8x7b-instruct"]
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Results CSV
|
||||
|
||||
`/opt/data/infra/assets/ai-optimizer/results.csv`
|
||||
|
||||
```csv
|
||||
timestamp,track,model,backend,phase,num_ctx,num_gpu,flash_attn,tokens_per_sec,vram_gb,ram_gb,status,is_best
|
||||
2026-04-28T17:00:00Z,gpu,devstral-small-2:24b,ollama,context_scaling,65536,99,true,15.2,52.1,18.4,success,false
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Cron Job Flow
|
||||
|
||||
```
|
||||
1. Read state.json
|
||||
2. If both queues empty → STOP (all models tested)
|
||||
3. Select next model from current track queue
|
||||
4. Pull model if needed (docker exec ollama ollama pull)
|
||||
5. Create Modelfile / llama.cpp config with current test params
|
||||
6. Run benchmark (both prompts)
|
||||
7. Measure: tokens/sec, VRAM (rocm-smi), RAM (free -h)
|
||||
8. If successful:
|
||||
- Increase context (next step)
|
||||
- Update current_config in state
|
||||
9. If OOM/error:
|
||||
- Record last good config as best_configs[track][model]
|
||||
- Move to next model in queue
|
||||
10. Update state.json
|
||||
11. Append to results.csv
|
||||
12. Git commit + push to /opt/data/infra
|
||||
13. Send Matrix notification if available, else silent
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Matrix Notification (Optional)
|
||||
|
||||
```python
|
||||
# If matrix credentials available in environment
|
||||
if os.getenv("MATRIX_HOME_SERVER") and os.getenv("MATRIX_ACCESS_TOKEN"):
|
||||
# Send completion notification
|
||||
# Room: !ai-optimizer:lazyworkhorse.net (or similar)
|
||||
pass
|
||||
# Else: silent, just commit
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Files to Create
|
||||
|
||||
```
|
||||
/opt/data/infra/assets/ai-optimizer/
|
||||
├── state.json # Current progress
|
||||
├── results.csv # All test results
|
||||
├── best_configs.json # Final best configs (human-readable)
|
||||
└── CRON_JOB_DRAFT.md # This file
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Notes
|
||||
|
||||
- **No num_parallel**: Removed to avoid limiting other settings
|
||||
- **Two tracks**: GPU (coding/speed) vs RAM (knowledge/context)
|
||||
- **Both backends**: Test ollama first, then llama.cpp if available
|
||||
- **Host tests**: rocm-smi must run on host or privileged container
|
||||
- **Deploy**: ai-worker has sudo for nh/nixos-rebuild, must ask user first
|
||||
1
assets/ai-optimizer/results.csv
Normal file
1
assets/ai-optimizer/results.csv
Normal file
@@ -0,0 +1 @@
|
||||
timestamp,track,model,backend,phase,num_ctx,num_gpu,flash_attn,tokens_per_sec,vram_gb,ram_gb,status,is_best
|
||||
|
21
assets/ai-optimizer/state.json
Normal file
21
assets/ai-optimizer/state.json
Normal file
@@ -0,0 +1,21 @@
|
||||
{
|
||||
"track": "gpu",
|
||||
"current_model": "devstral-small-2:24b",
|
||||
"model_index": 0,
|
||||
"phase": "context_scaling",
|
||||
"backend": "ollama",
|
||||
"current_config": {
|
||||
"num_ctx": 32768,
|
||||
"num_gpu": 99,
|
||||
"flash_attn": true
|
||||
},
|
||||
"best_configs": {
|
||||
"gpu": {},
|
||||
"ram": {}
|
||||
},
|
||||
"completed_models": [],
|
||||
"gpu_queue": ["devstral-small-2:24b", "qwen2.5-coder:32b", "codellama:34b-instruct"],
|
||||
"ram_queue": ["qwen2.5:72b", "nemotron-3-nano:30b", "mixtral:8x7b-instruct"],
|
||||
"context_steps": [32768, 65536, 98304, 131072, 163840, 200704, 262144, 327680],
|
||||
"last_updated": "2026-04-28T17:00:00Z"
|
||||
}
|
||||
Submodule assets/compose updated: fb0f2cbe84...ba8634b770
21
docker/hermes/Dockerfile
Normal file
21
docker/hermes/Dockerfile
Normal file
@@ -0,0 +1,21 @@
|
||||
# Custom Hermes Agent image with voice support and ROCm for AMD GPU
|
||||
# Builds on upstream image and adds local STT capabilities
|
||||
FROM nousresearch/hermes-agent:latest AS base
|
||||
|
||||
# Install system dependencies for audio capture
|
||||
USER root
|
||||
RUN apt-get update && \
|
||||
apt-get install -y --no-install-recommends \
|
||||
libportaudio2 \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Install Python packages for voice support
|
||||
# Using uv for consistency with Hermes build process
|
||||
RUN uv pip install --no-cache-dir \
|
||||
sounddevice \
|
||||
numpy \
|
||||
faster-whisper
|
||||
|
||||
# Health check for voice readiness
|
||||
HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
|
||||
CMD python3 -c "import sounddevice, faster_whisper; print('voice deps OK')" || exit 1
|
||||
102
docker/hermes/entrypoint.sh
Executable file
102
docker/hermes/entrypoint.sh
Executable file
@@ -0,0 +1,102 @@
|
||||
#!/bin/bash
|
||||
# Docker/Podman entrypoint: bootstrap config files into the mounted volume, then run hermes.
|
||||
set -e
|
||||
|
||||
HERMES_HOME="${HERMES_HOME:-/opt/data}"
|
||||
INSTALL_DIR="/opt/hermes"
|
||||
|
||||
# --- Privilege dropping via gosu ---
|
||||
# When started as root (the default for Docker, or fakeroot in rootless Podman),
|
||||
# optionally remap the hermes user/group to match host-side ownership, fix volume
|
||||
# permissions, then re-exec as hermes.
|
||||
if [ "$(id -u)" = "0" ]; then
|
||||
if [ -n "$HERMES_UID" ] && [ "$HERMES_UID" != "$(id -u hermes)" ]; then
|
||||
echo "Changing hermes UID to $HERMES_UID"
|
||||
usermod -u "$HERMES_UID" hermes
|
||||
fi
|
||||
|
||||
if [ -n "$HERMES_GID" ] && [ "$HERMES_GID" != "$(id -g hermes)" ]; then
|
||||
echo "Changing hermes GID to $HERMES_GID"
|
||||
# -o allows non-unique GID (e.g. macOS GID 20 "staff" may already exist
|
||||
# as "dialout" in the Debian-based container image)
|
||||
groupmod -o -g "$HERMES_GID" hermes 2>/dev/null || true
|
||||
fi
|
||||
|
||||
# Fix ownership of the data volume. When HERMES_UID remaps the hermes user,
|
||||
# files created by previous runs (under the old UID) become inaccessible.
|
||||
# Always chown -R when UID was remapped; otherwise only if top-level is wrong.
|
||||
actual_hermes_uid=$(id -u hermes)
|
||||
needs_chown=false
|
||||
if [ -n "$HERMES_UID" ] && [ "$HERMES_UID" != "10000" ]; then
|
||||
needs_chown=true
|
||||
elif [ "$(stat -c %u "$HERMES_HOME" 2>/dev/null)" != "$actual_hermes_uid" ]; then
|
||||
needs_chown=true
|
||||
fi
|
||||
if [ "$needs_chown" = true ]; then
|
||||
echo "Fixing ownership of $HERMES_HOME to hermes ($actual_hermes_uid)"
|
||||
# In rootless Podman the container's "root" is mapped to an unprivileged
|
||||
# host UID — chown will fail. That's fine: the volume is already owned
|
||||
# by the mapped user on the host side.
|
||||
chown -R hermes:hermes "$HERMES_HOME" 2>/dev/null || \
|
||||
echo "Warning: chown failed (rootless container?) — continuing anyway"
|
||||
fi
|
||||
|
||||
echo "Dropping root privileges"
|
||||
exec gosu hermes "$0" "$@"
|
||||
fi
|
||||
|
||||
# --- Running as hermes from here ---
|
||||
source "${INSTALL_DIR}/.venv/bin/activate"
|
||||
|
||||
# Create essential directory structure. Cache and platform directories
|
||||
# (cache/images, cache/audio, platforms/whatsapp, etc.) are created on
|
||||
# demand by the application — don't pre-create them here so new installs
|
||||
# get the consolidated layout from get_hermes_dir().
|
||||
# The "home/" subdirectory is a per-profile HOME for subprocesses (git,
|
||||
# ssh, gh, npm …). Without it those tools write to /root which is
|
||||
# ephemeral and shared across profiles. See issue #4426.
|
||||
mkdir -p "$HERMES_HOME"/{cron,sessions,logs,hooks,memories,skills,skins,plans,workspace,home}
|
||||
|
||||
# .env
|
||||
if [ ! -f "$HERMES_HOME/.env" ]; then
|
||||
cp "$INSTALL_DIR/.env.example" "$HERMES_HOME/.env"
|
||||
fi
|
||||
|
||||
# config.yaml
|
||||
if [ ! -f "$HERMES_HOME/config.yaml" ]; then
|
||||
cp "$INSTALL_DIR/cli-config.yaml.example" "$HERMES_HOME/config.yaml"
|
||||
fi
|
||||
|
||||
# Ensure the main config file remains accessible to the hermes runtime user
|
||||
# even if it was edited on the host after initial ownership setup.
|
||||
if [ -f "$HERMES_HOME/config.yaml" ]; then
|
||||
chown hermes:hermes "$HERMES_HOME/config.yaml"
|
||||
chmod 640 "$HERMES_HOME/config.yaml"
|
||||
fi
|
||||
|
||||
# SOUL.md
|
||||
if [ ! -f "$HERMES_HOME/SOUL.md" ]; then
|
||||
cp "$INSTALL_DIR/docker/SOUL.md" "$HERMES_HOME/SOUL.md"
|
||||
fi
|
||||
|
||||
# Sync bundled skills (manifest-based so user edits are preserved)
|
||||
if [ -d "$INSTALL_DIR/skills" ]; then
|
||||
python3 "$INSTALL_DIR/tools/skills_sync.py"
|
||||
fi
|
||||
|
||||
# Final exec: two supported invocation patterns.
|
||||
#
|
||||
# docker run <image> -> exec `hermes` with no args (legacy default)
|
||||
# docker run <image> chat -q "..." -> exec `hermes chat -q "..."` (legacy wrap)
|
||||
# docker run <image> sleep infinity -> exec `sleep infinity` directly
|
||||
# docker run <image> bash -> exec `bash` directly
|
||||
#
|
||||
# If the first positional arg resolves to an executable on PATH, we assume the
|
||||
# caller wants to run it directly (needed by the launcher which runs long-lived
|
||||
# `sleep infinity` sandbox containers — see tools/environments/docker.py).
|
||||
# Otherwise we treat the args as a hermes subcommand and wrap with `hermes`,
|
||||
# preserving the documented `docker run <image> <subcommand>` behavior.
|
||||
if [ $# -gt 0 ] && command -v "$1" >/dev/null 2>&1; then
|
||||
exec "$@"
|
||||
fi
|
||||
exec hermes "$@"
|
||||
@@ -1,474 +0,0 @@
|
||||
# Nix Installation for Hermes Agent Container
|
||||
|
||||
This guide covers several approaches for installing Nix in the Hermes Agent Docker
|
||||
container to enable remote NixOS deployment via `nixos-rebuild`. It covers both
|
||||
x86_64 (lazyworkhorse) and aarch64 (cyt-pi, uConsole) architectures.
|
||||
|
||||
## Table of Contents
|
||||
|
||||
1. [Why Nix in a Container?](#why-nix-in-a-container)
|
||||
2. [Prerequisites](#prerequisites)
|
||||
3. [Installation Methods](#installation-methods)
|
||||
- [Method A: Determinate Systems Installer](#method-a-determinate-systems-installer-recommended)
|
||||
- [Method B: Vanilla Nix Installer](#method-b-vanilla-nix-installer)
|
||||
- [Method C: NixOS-Based Container Image](#method-c-nixos-based-container-image)
|
||||
4. [Architecture-Specific Notes](#architecture-specific-notes)
|
||||
- [x86_64 (lazyworkhorse)](#x86_64-lazyworkhorse)
|
||||
- [aarch64 (cyt-pi, uConsole)](#aarch64-cyt-pi-uconsole)
|
||||
- [Cross-Compilation](#cross-compilation)
|
||||
5. [Post-Install Configuration](#post-install-configuration)
|
||||
6. [Verification](#verification)
|
||||
7. [Container-Specific Considerations](#container-specific-considerations)
|
||||
- [Persistence](#persistence)
|
||||
- [Disk Space](#disk-space)
|
||||
- [Security](#security)
|
||||
- [Resource Constraints](#resource-constraints)
|
||||
8. [Integration with deploy.sh](#integration-with-deploysh)
|
||||
9. [Troubleshooting](#troubleshooting)
|
||||
10. [References](#references)
|
||||
|
||||
---
|
||||
|
||||
## Why Nix in a Container?
|
||||
|
||||
The Hermes Agent container runs on an Ubuntu/Debian base. To deploy NixOS
|
||||
configurations to remote hosts, we need:
|
||||
|
||||
- `nix` — the Nix package manager (for building configurations)
|
||||
- `nixos-rebuild` — the NixOS deployment tool
|
||||
- Access to the infra repo with flake configuration
|
||||
|
||||
Installing Nix inside the container avoids:
|
||||
- Host-level Nix installation on the Docker host
|
||||
- Cross-container volume mounts of /nix/store
|
||||
- Dependencies on the host's Nix daemon (which may be a different version)
|
||||
|
||||
## Prerequisites
|
||||
|
||||
- Docker host running Linux (x86_64 and/or aarch64)
|
||||
- Container base: Debian/Ubuntu (apt-based)
|
||||
- 1-2 GB additional disk space for Nix store
|
||||
- Network access to cache.nixos.org (or a local binary cache)
|
||||
- Git access to the infra repository
|
||||
|
||||
## Installation Methods
|
||||
|
||||
### Method A: Determinate Systems Installer (Recommended)
|
||||
|
||||
The Determinate Systems installer is the recommended approach. It is non-interactive,
|
||||
sets up flakes by default, and handles multi-user installation cleanly.
|
||||
|
||||
**Dockerfile additions:**
|
||||
|
||||
```dockerfile
|
||||
# Install Nix (Determinate Systems installer)
|
||||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||||
curl \
|
||||
xz-utils \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Download and run Nix installer (non-interactive)
|
||||
RUN curl --proto '=https' --tlsv1.2 -sSf -L https://install.determinate.systems/nix \
|
||||
-o /tmp/nix-install.sh \
|
||||
&& chmod +x /tmp/nix-install.sh \
|
||||
&& sh /tmp/nix-install.sh install --no-confirm \
|
||||
&& rm /tmp/nix-install.sh
|
||||
|
||||
# Configure Nix for flakes
|
||||
RUN mkdir -p /root/.config/nix \
|
||||
&& echo 'experimental-features = nix-command flakes' > /root/.config/nix/nix.conf
|
||||
|
||||
# Add Nix to PATH for all users
|
||||
ENV PATH="/nix/var/nix/profiles/default/bin:$PATH"
|
||||
```
|
||||
|
||||
**Pros:**
|
||||
- Fully non-interactive (--no-confirm)
|
||||
- Enables flakes automatically
|
||||
- Sets up multi-user daemon
|
||||
- Auto-selects correct architecture
|
||||
- Handles upgrades gracefully
|
||||
|
||||
**Cons:**
|
||||
- Downloads ~100 MB installer
|
||||
- Requires systemd in container (works with --privileged or cgroupv2)
|
||||
- Daemon mode may conflict with container exit semantics
|
||||
|
||||
**Container runtime additions:**
|
||||
|
||||
For the Nix daemon to work properly inside a container, you may need:
|
||||
```dockerfile
|
||||
# Ensure /nix is a volume for persistence
|
||||
VOLUME /nix
|
||||
|
||||
# Or mount tmpfs for ephemeral builds:
|
||||
# docker run --tmpfs /nix:exec,size=4G ...
|
||||
```
|
||||
|
||||
### Method B: Vanilla Nix Installer
|
||||
|
||||
The official single-user Nix installer is lighter but requires manual flake setup.
|
||||
|
||||
**Dockerfile additions:**
|
||||
|
||||
```dockerfile
|
||||
# Install Nix (single-user, official installer)
|
||||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||||
curl \
|
||||
sudo \
|
||||
xz-utils \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Install Nix as root (single-user)
|
||||
RUN curl -L https://nixos.org/nix/install -o /tmp/nix-install.sh \
|
||||
&& chmod +x /tmp/nix-install.sh \
|
||||
&& sh /tmp/nix-install.sh --no-daemon \
|
||||
&& rm /tmp/nix-install.sh
|
||||
|
||||
# Enable flakes
|
||||
RUN mkdir -p /root/.config/nix \
|
||||
&& echo 'experimental-features = nix-command flakes' > /root/.config/nix/nix.conf
|
||||
|
||||
# Source Nix in shell
|
||||
RUN echo '. /root/.nix-profile/etc/profile.d/nix.sh' >> /root/.bashrc
|
||||
ENV PATH="/root/.nix-profile/bin:$PATH"
|
||||
```
|
||||
|
||||
**Pros:**
|
||||
- Smaller installer
|
||||
- No daemon needed (single-user mode)
|
||||
- Works in containers without systemd
|
||||
- Simpler container lifecycle
|
||||
|
||||
**Cons:**
|
||||
- Manual flake configuration required
|
||||
- Single-user only (no multi-user isolation)
|
||||
- PATH must be set manually
|
||||
- No automatic garbage collection
|
||||
|
||||
### Method C: NixOS-Based Container Image
|
||||
|
||||
For maximum isolation, use an official NixOS base image for the build stage.
|
||||
|
||||
**Multi-stage Dockerfile:**
|
||||
|
||||
```dockerfile
|
||||
# Build stage: NixOS builder
|
||||
FROM nixos/nix:latest AS builder
|
||||
|
||||
COPY infra /infra
|
||||
WORKDIR /infra
|
||||
|
||||
# Build the configuration once
|
||||
RUN nix build '.#nixosConfigurations.lazyworkhorse.config.system.build.toplevel'
|
||||
|
||||
# Final stage: Hermes container
|
||||
FROM ubuntu:22.04
|
||||
|
||||
# Copy the Nix closure and binary cache
|
||||
COPY --from=builder /nix /nix
|
||||
|
||||
# ... rest of Hermes setup
|
||||
```
|
||||
|
||||
**Pros:**
|
||||
- Purely declarative build environment
|
||||
- No installation at runtime
|
||||
- Easy to pin Nix version
|
||||
- Good for CI/CD pipelines
|
||||
|
||||
**Cons:**
|
||||
- Requires multi-stage Docker build
|
||||
- Larger initial image build
|
||||
- Harder to update Nix version at runtime
|
||||
- Overkill if Nix is only needed for `nixos-rebuild`
|
||||
|
||||
---
|
||||
|
||||
## Architecture-Specific Notes
|
||||
|
||||
### x86_64 (lazyworkhorse)
|
||||
|
||||
The Hermes container likely runs on x86_64 hardware for the primary server.
|
||||
Nix will download x86_64 binaries from cache.nixos.org by default.
|
||||
|
||||
**No special configuration needed** — the standard installer works out of the box.
|
||||
|
||||
If the container is running on an AMD Ryzen/EPYC or Intel Xeon, consider:
|
||||
```bash
|
||||
# Enable CPU-specific optimizations (optional)
|
||||
echo 'extra-platforms = x86_64-v1 x86_64-v2 x86_64-v3' >> /root/.config/nix/nix.conf
|
||||
```
|
||||
|
||||
### aarch64 (cyt-pi, uConsole)
|
||||
|
||||
When building for aarch64 targets from an x86_64 container, you need either:
|
||||
1. Remote builder (aarch64 machine does the build), or
|
||||
2. QEMU-based emulation (slower but self-contained), or
|
||||
3. Build directly on the aarch64 target using `--build-host`
|
||||
|
||||
**For QEMU emulation in the container:**
|
||||
|
||||
```dockerfile
|
||||
# Enable binfmt for aarch64 emulation
|
||||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||||
qemu-user-static \
|
||||
binfmt-support \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Register aarch64 binfmt
|
||||
RUN update-binfmts --enable qemu-aarch64
|
||||
```
|
||||
|
||||
**Container runtime (for QEMU):**
|
||||
```bash
|
||||
docker run --privileged --rm ... hermes-agent
|
||||
# Or with specific capability:
|
||||
docker run --cap-add=SYS_ADMIN --security-opt seccomp=unconfined ... hermes-agent
|
||||
```
|
||||
|
||||
### Cross-Compilation
|
||||
|
||||
For native cross-compilation (without emulation), add to your Nix configuration:
|
||||
|
||||
```nix
|
||||
# In your flake.nix or nix.conf
|
||||
{
|
||||
nix.settings.extra-platforms = [ "aarch64-linux" "x86_64-linux" ];
|
||||
nix.settings.extra-sandbox-paths = [ ];
|
||||
boot.binfmt.emulatedSystems = [ "aarch64-linux" ];
|
||||
}
|
||||
```
|
||||
|
||||
Or in `nix.conf`:
|
||||
```
|
||||
extra-platforms = x86_64-linux aarch64-linux
|
||||
extra-sandbox-paths =
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Post-Install Configuration
|
||||
|
||||
### nix.conf for Container Usage
|
||||
|
||||
Recommended `/root/.config/nix/nix.conf`:
|
||||
|
||||
```ini
|
||||
experimental-features = nix-command flakes
|
||||
substituters = https://cache.nixos.org/
|
||||
trusted-users = root
|
||||
max-jobs = auto
|
||||
cores = 0
|
||||
sandbox = false
|
||||
```
|
||||
|
||||
Note: `sandbox = false` is needed inside containers that lack full sandbox
|
||||
support. This is safe in a single-tenant container environment.
|
||||
|
||||
### PATH Setup
|
||||
|
||||
Add to your Dockerfile:
|
||||
```dockerfile
|
||||
ENV PATH="/nix/var/nix/profiles/default/bin:/root/.nix-profile/bin:${PATH}"
|
||||
```
|
||||
|
||||
### Shell Integration
|
||||
|
||||
```dockerfile
|
||||
RUN echo 'source /root/.nix-profile/etc/profile.d/nix.sh' >> /root/.bashrc
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Verification
|
||||
|
||||
After installation, verify with:
|
||||
|
||||
```bash
|
||||
# Check Nix is available
|
||||
nix --version
|
||||
|
||||
# Check nixos-rebuild
|
||||
nixos-rebuild --help | head -3
|
||||
|
||||
# Verify flakes are enabled
|
||||
nix flake --help
|
||||
|
||||
# Test a build (must be in infra repo)
|
||||
cd /opt/data/infra
|
||||
nix build --no-link '.#nixosConfigurations.lazyworkhorse.config.system.build.toplevel'
|
||||
|
||||
# Check available systems
|
||||
nix eval --impure --expr 'builtins.currentSystem'
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Container-Specific Considerations
|
||||
|
||||
### Persistence
|
||||
|
||||
The `/nix` directory should be a Docker volume to avoid re-downloading
|
||||
packages on every container restart:
|
||||
|
||||
```yaml
|
||||
# docker-compose.yml
|
||||
volumes:
|
||||
- nix-store:/nix
|
||||
|
||||
volumes:
|
||||
nix-store:
|
||||
```
|
||||
|
||||
Without persistence, every container restart requires re-downloading the
|
||||
entire Nix store (~500 MB - 2 GB depending on packages used).
|
||||
|
||||
### Disk Space
|
||||
|
||||
The Nix store grows over time as old generations accumulate. Set up garbage
|
||||
collection:
|
||||
|
||||
```bash
|
||||
# Manual GC
|
||||
nix store gc
|
||||
|
||||
# Remove old generations
|
||||
nix-collect-garbage --delete-older-than 30d
|
||||
|
||||
# Automatic GC (in nix.conf)
|
||||
# Currently not supported in nix.conf, but you can run a cron job:
|
||||
# nix store gc --max 10G
|
||||
```
|
||||
|
||||
In Docker, limit store growth with:
|
||||
```dockerfile
|
||||
# Configure max store size
|
||||
RUN mkdir -p /etc/nix && \
|
||||
echo 'min-free = 5368709120' > /etc/nix/nix.conf # Keep 5GB free
|
||||
```
|
||||
|
||||
### Security
|
||||
|
||||
Running Nix in a container introduces some security considerations:
|
||||
|
||||
1. **Sandboxing:** `sandbox = false` disables build isolation. In a multi-tenant
|
||||
container, this means Nix builds can affect the container filesystem.
|
||||
**Mitigation:** Only build configs you trust (your own infra repo).
|
||||
|
||||
2. **Network access:** The container needs outbound access to cache.nixos.org.
|
||||
If using a restricted network, set up a local binary cache:
|
||||
```nix
|
||||
substituters = https://cache.nixos.org/ https://nix-cache.internal/
|
||||
```
|
||||
|
||||
3. **Privileged mode:** QEMU emulation for aarch64 builds may need `--privileged`
|
||||
or `--security-opt seccomp=unconfined`. This reduces container isolation.
|
||||
**Mitigation:** Use remote builders or build natively on the target.
|
||||
|
||||
4. **Supply chain:** Nix derivations pin exact inputs via hashes. Verify
|
||||
flake.lock is committed and reviewed.
|
||||
|
||||
### Resource Constraints
|
||||
|
||||
Nix builds can be memory and CPU intensive:
|
||||
|
||||
```nix
|
||||
# Limit build parallelism in nix.conf
|
||||
max-jobs = 2
|
||||
cores = 4
|
||||
|
||||
# Or set per-build:
|
||||
# nix build --max-jobs 2 --cores 4
|
||||
```
|
||||
|
||||
For containers with limited memory (< 2 GB), consider:
|
||||
- Building on the target host instead (`--build-host`)
|
||||
- Using the deploy script's `build` action separately
|
||||
|
||||
---
|
||||
|
||||
## Integration with deploy.sh
|
||||
|
||||
The deployment script at `scripts/deploy.sh` expects:
|
||||
|
||||
1. **Nix installed** with flakes enabled
|
||||
2. **SSH key** at `/opt/data/home/.ssh/id_hermes_gitea` (or via SSH_KEY env)
|
||||
3. **Infra repo** cloned at the script's parent directory
|
||||
4. **Network access** to:
|
||||
- `code.lazyworkhorse.net:2222` (Gitea for git operations)
|
||||
- Target hosts via SSH (see deploy-ssh-config)
|
||||
- `cache.nixos.org` or a local substitute
|
||||
|
||||
Typical usage from Hermes:
|
||||
|
||||
```bash
|
||||
# Full deployment
|
||||
./scripts/deploy.sh lazyworkhorse master switch
|
||||
|
||||
# Build-only check (no remote deployment)
|
||||
./scripts/deploy.sh cyt-pi master build
|
||||
|
||||
# Dry run
|
||||
./scripts/deploy.sh uConsole feat/test dry-activate
|
||||
|
||||
# Override SSH key
|
||||
SSH_KEY=/opt/data/home/.ssh/my-custom-key ./deploy.sh lazyworkhorse
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Troubleshooting
|
||||
|
||||
### "nix: command not found"
|
||||
|
||||
- Ensure Nix is installed and PATH is set:
|
||||
```bash
|
||||
export PATH="/nix/var/nix/profiles/default/bin:/root/.nix-profile/bin:$PATH"
|
||||
```
|
||||
- Check installation: `ls -la /nix/` should exist
|
||||
- Re-source profile: `. /root/.nix-profile/etc/profile.d/nix.sh`
|
||||
|
||||
### "error: unable to download ... cache.nixos.org"
|
||||
|
||||
- Check network connectivity: `ping cache.nixos.org`
|
||||
- Check DNS resolution from inside the container
|
||||
- If behind a proxy, set `http_proxy` / `https_proxy` environment variables
|
||||
|
||||
### "sandbox: cannot run build in sandbox"
|
||||
|
||||
- Add `sandbox = false` to nix.conf
|
||||
- Or run container with `--privileged` or `--security-opt seccomp=unconfined`
|
||||
|
||||
### "aarch64-linux builds fail on x86_64"
|
||||
|
||||
- QEMU binfmt not registered. Check: `ls /proc/sys/fs/binfmt_misc/`
|
||||
- Rebuild QEMU registration: `docker run --privileged --rm tonistiigi/binfmt --install all`
|
||||
- Or use `--build-host` to build on the target directly
|
||||
|
||||
### "nixos-rebuild fails with SSH errors"
|
||||
|
||||
- Verify SSH key exists and has correct permissions:
|
||||
```bash
|
||||
ls -la /opt/data/home/.ssh/id_hermes_gitea
|
||||
chmod 600 /opt/data/home/.ssh/id_hermes_gitea
|
||||
```
|
||||
- Test SSH manually: `ssh -p 2424 -i /opt/data/home/.ssh/id_hermes_gitea ai-worker@lazyworkhorse.net`
|
||||
- Check target host is reachable: `nc -zv lazyworkhorse.net 2424`
|
||||
|
||||
### "git fetch fails from Gitea"
|
||||
|
||||
- Verify GIT_SSH_COMMAND is set: `echo $GIT_SSH_COMMAND`
|
||||
- Test git SSH: `ssh -T git@code.lazyworkhorse.net -p 2222`
|
||||
- Check the infra repo remote: `git remote -v`
|
||||
|
||||
---
|
||||
|
||||
## References
|
||||
|
||||
- [Determinate Systems Nix Installer](https://github.com/DeterminateSystems/nix-installer)
|
||||
- [NixOS Manual: Installation](https://nixos.org/manual/nix/stable/installation/)
|
||||
- [NixOS Wiki: Flakes](https://nixos.wiki/wiki/Flakes)
|
||||
- [NixOS Wiki: nixos-rebuild](https://nixos.wiki/wiki/Nixos-rebuild)
|
||||
- [NixOS Wiki: Cross Compilation](https://nixos.wiki/wiki/Cross_Compilation)
|
||||
- [Multi-arch Docker with QEMU](https://github.com/multiarch/qemu-user-static)
|
||||
@@ -1,63 +0,0 @@
|
||||
# Hermes Container SSH Configuration
|
||||
# For NixOS deployment to remote hosts
|
||||
#
|
||||
# Usage:
|
||||
# cp scripts/deploy-ssh-config ~/.ssh/config.d/hermes-include
|
||||
# Or: cat scripts/deploy-ssh-config >> ~/.ssh/config
|
||||
#
|
||||
# This config covers all NixOS hosts managed from the Hermes container.
|
||||
# Lazyworkhorse has two users: ai-worker (primary automation) and gortium (admin).
|
||||
# Cyt-pi connects via reverse SSH tunnel on port 19999.
|
||||
# uConsole is a placeholder until LAN-hostname resolution is confirmed.
|
||||
|
||||
# ── Global defaults ──────────────────────────────────────────────────
|
||||
Host *
|
||||
ServerAliveInterval 60
|
||||
ServerAliveCountMax 3
|
||||
TCPKeepAlive yes
|
||||
Compression yes
|
||||
CompressionLevel 6
|
||||
ControlMaster auto
|
||||
ControlPath ~/.ssh/controlmasters/%r@%h:%p
|
||||
ControlPersist 10m
|
||||
StrictHostKeyChecking no
|
||||
UserKnownHostsFile /dev/null
|
||||
|
||||
# ── Hosts ──────────────────────────────────────────────────────────────
|
||||
|
||||
# Lazyworkhorse — x86_64 main server (ai-worker@lazyworkhorse.net:2424)
|
||||
Host lazyworkhorse
|
||||
HostName lazyworkhorse.net
|
||||
User ai-worker
|
||||
Port 2424
|
||||
IdentityFile /opt/data/home/.ssh/id_hermes_gitea
|
||||
|
||||
# Lazyworkhorse — admin access (gortium@lazyworkhorse.net:2425)
|
||||
Host lazyworkhorse-admin
|
||||
HostName lazyworkhorse.net
|
||||
User gortium
|
||||
Port 2425
|
||||
IdentityFile /opt/data/home/.ssh/id_hermes_gitea
|
||||
|
||||
# Cyt-pi — aarch64 Pi Zero 2 W
|
||||
# Connected via reverse SSH tunnel (gortium directs tunnel to :19999)
|
||||
Host cyt-pi
|
||||
HostName localhost
|
||||
User gortium
|
||||
Port 19999
|
||||
IdentityFile /opt/data/home/.ssh/id_hermes_gitea
|
||||
|
||||
# uConsole — aarch64 ClockworkPi (placeholder hostname)
|
||||
# Replace uconsole.lan with actual IP/hostname when deployed
|
||||
Host uConsole uconsole
|
||||
HostName uconsole.lan
|
||||
User gortium
|
||||
Port 22
|
||||
IdentityFile /opt/data/home/.ssh/id_hermes_gitea
|
||||
|
||||
# ── Gitea host — for git operations ──────────────────────────────────
|
||||
Host code
|
||||
HostName code.lazyworkhorse.net
|
||||
Port 2222
|
||||
User gortium
|
||||
IdentityFile /opt/data/home/.ssh/id_hermes_gitea
|
||||
@@ -1,286 +0,0 @@
|
||||
#!/usr/bin/env bash
|
||||
# NixOS Deployment Helper Script
|
||||
# Remote NixOS deployment from Hermes container to target hosts.
|
||||
#
|
||||
# Usage: ./deploy.sh <hostname> [branch] [action]
|
||||
#
|
||||
# Actions:
|
||||
# switch Activate configuration now (default)
|
||||
# boot Activate on next reboot
|
||||
# test Activate without switching generations
|
||||
# build Build locally only, no remote activation
|
||||
# dry-activate Show what would change without applying
|
||||
#
|
||||
# Examples:
|
||||
# ./deploy.sh lazyworkhorse # deploy master/switch to lazyworkhorse
|
||||
# ./deploy.sh cyt-pi feat/test boot # deploy feat/test branch, activate on boot
|
||||
# ./deploy.sh uConsole master build # just build, don't deploy
|
||||
# NO_BUILD_CHECK=1 ./deploy.sh uConsole # skip the pre-flight nix build
|
||||
#
|
||||
# Environment variables:
|
||||
# SSH_USER SSH user (default: auto-detected per host)
|
||||
# SSH_PORT SSH port (default: auto-detected per host)
|
||||
# SSH_KEY SSH identity file
|
||||
# BUILD_HOST Build flake for this host (default: same as target host)
|
||||
# NO_BUILD_CHECK Set to 1 to skip local nix build before deployment
|
||||
|
||||
set -euo pipefail
|
||||
|
||||
# ── Colors ──────────────────────────────────────────────────────────────
|
||||
RED='\033[0;31m'
|
||||
GREEN='\033[0;32m'
|
||||
YELLOW='\033[1;33m'
|
||||
BLUE='\033[0;34m'
|
||||
CYAN='\033[0;36m'
|
||||
NC='\033[0m' # No Color
|
||||
|
||||
info() { echo -e "${BLUE}[INFO]${NC} $*"; }
|
||||
ok() { echo -e "${GREEN}[OK]${NC} $*"; }
|
||||
warn() { echo -e "${YELLOW}[WARN]${NC} $*"; }
|
||||
error() { echo -e "${RED}[ERROR]${NC} $*" >&2; }
|
||||
step() { echo -e "\n${CYAN}━━━ $* ━━━${NC}"; }
|
||||
|
||||
# ── Cleanup trap ───────────────────────────────────────────────────────
|
||||
cleanup() {
|
||||
local ec=$?
|
||||
if [ $ec -ne 0 ]; then
|
||||
error "Deployment failed with exit code $ec"
|
||||
fi
|
||||
exit $ec
|
||||
}
|
||||
trap cleanup EXIT
|
||||
|
||||
# ── Usage / Help ───────────────────────────────────────────────────────
|
||||
show_usage() {
|
||||
cat <<EOF
|
||||
Usage: $0 <hostname> [branch] [action]
|
||||
|
||||
Remote NixOS deployment from Hermes container to target hosts.
|
||||
|
||||
HOSTNAME (required):
|
||||
lazyworkhorse x86_64 main server
|
||||
cyt-pi aarch64 Pi Zero 2 W (via reverse tunnel)
|
||||
uConsole aarch64 ClockworkPi
|
||||
|
||||
BRANCH (optional, default: master):
|
||||
Git branch or tag to deploy. Fetched from origin.
|
||||
|
||||
ACTION (optional, default: switch):
|
||||
switch Activate configuration now (default)
|
||||
boot Activate on next reboot
|
||||
test Activate without switching generations
|
||||
build Build locally only, skip remote deployment
|
||||
dry-activate Show what would change without applying
|
||||
|
||||
Environment variables:
|
||||
SSH_USER SSH username override
|
||||
SSH_PORT SSH port override
|
||||
SSH_KEY SSH identity file path
|
||||
BUILD_HOST Build flake hostname (default: same as HOSTNAME)
|
||||
NO_BUILD_CHECK Skip local nix build validation (set to 1)
|
||||
|
||||
Examples:
|
||||
$0 lazyworkhorse # deploy master/switch
|
||||
$0 cyt-pi feat/test boot # deploy feature branch, boot
|
||||
$0 uConsole master build # just build, no remote
|
||||
NO_BUILD_CHECK=1 $0 uConsole # skip build check
|
||||
|
||||
EOF
|
||||
exit 0
|
||||
}
|
||||
|
||||
# ── Argument parsing ───────────────────────────────────────────────────
|
||||
HOSTNAME="${1:-}"
|
||||
BRANCH="${2:-master}"
|
||||
ACTION="${3:-switch}"
|
||||
NO_BUILD_CHECK="${NO_BUILD_CHECK:-0}"
|
||||
|
||||
if [ "$HOSTNAME" = "--help" ] || [ "$HOSTNAME" = "-h" ] || [ -z "$HOSTNAME" ]; then
|
||||
show_usage
|
||||
fi
|
||||
|
||||
# ── Host configuration ─────────────────────────────────────────────────
|
||||
case "$HOSTNAME" in
|
||||
lazyworkhorse)
|
||||
DEFAULT_SSH_USER="ai-worker"
|
||||
DEFAULT_SSH_PORT="2424"
|
||||
ARCH="x86_64-linux"
|
||||
;;
|
||||
cyt-pi)
|
||||
DEFAULT_SSH_USER="gortium"
|
||||
DEFAULT_SSH_PORT="19999"
|
||||
ARCH="aarch64-linux"
|
||||
;;
|
||||
uConsole)
|
||||
DEFAULT_SSH_USER="gortium"
|
||||
DEFAULT_SSH_PORT="22"
|
||||
ARCH="aarch64-linux"
|
||||
;;
|
||||
*)
|
||||
error "Unknown host: $HOSTNAME"
|
||||
echo "Supported hosts: lazyworkhorse, cyt-pi, uConsole"
|
||||
exit 1
|
||||
;;
|
||||
esac
|
||||
|
||||
SSH_USER="${SSH_USER:-$DEFAULT_SSH_USER}"
|
||||
SSH_PORT="${SSH_PORT:-$DEFAULT_SSH_PORT}"
|
||||
SSH_KEY="${SSH_KEY:-/opt/data/home/.ssh/id_hermes_gitea}"
|
||||
BUILD_HOST="${BUILD_HOST:-$HOSTNAME}"
|
||||
|
||||
SSH_OPTS="-p $SSH_PORT -i $SSH_KEY -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null"
|
||||
SSH_TARGET="${SSH_USER}@${HOSTNAME}"
|
||||
export GIT_SSH_COMMAND="ssh -i $SSH_KEY -p 2222 -o StrictHostKeyChecking=no"
|
||||
export PATH="/nix/var/nix/profiles/default/bin:$PATH"
|
||||
|
||||
# ── Banner ─────────────────────────────────────────────────────────────
|
||||
echo "╔══════════════════════════════════════════════╗"
|
||||
echo "║ NixOS Remote Deployment ║"
|
||||
echo "╚══════════════════════════════════════════════╝"
|
||||
info "Host: $HOSTNAME ($ARCH)"
|
||||
info "Branch: $BRANCH"
|
||||
info "Action: $ACTION"
|
||||
info "SSH: ${SSH_USER}@${HOSTNAME}:${SSH_PORT}"
|
||||
echo ""
|
||||
|
||||
# ── Pre-flight checks ─────────────────────────────────────────────────
|
||||
step "Pre-flight checks"
|
||||
|
||||
# 1. Check required tools
|
||||
for cmd in nix git ssh; do
|
||||
if ! command -v "$cmd" &>/dev/null; then
|
||||
error "Required tool not found: $cmd"
|
||||
exit 1
|
||||
fi
|
||||
done
|
||||
ok "Required tools available (nix, git, ssh)"
|
||||
|
||||
# 2. Check infra repo
|
||||
INFRA_DIR="$(cd "$(dirname "$0")/.." && pwd)"
|
||||
if [ ! -d "$INFRA_DIR/.git" ]; then
|
||||
error "Not a git repository: $INFRA_DIR"
|
||||
exit 1
|
||||
fi
|
||||
ok "Infra repo found at $INFRA_DIR"
|
||||
|
||||
# 3. Check SSH connectivity (skip for build-only actions)
|
||||
if [ "$ACTION" != "build" ]; then
|
||||
if ssh $SSH_OPTS -o ConnectTimeout=5 "$SSH_TARGET" "echo connected" &>/dev/null; then
|
||||
ok "SSH connectivity to $HOSTNAME verified"
|
||||
else
|
||||
warn "Cannot reach $HOSTNAME via SSH — deployment step will fail later"
|
||||
fi
|
||||
fi
|
||||
|
||||
# ── Git sync ───────────────────────────────────────────────────────────
|
||||
step "Git sync"
|
||||
|
||||
cd "$INFRA_DIR"
|
||||
|
||||
# Stash local changes if any
|
||||
if ! git diff --quiet HEAD; then
|
||||
warn "Local changes detected, stashing..."
|
||||
git stash push -m "auto-stash before deploy $(date -Iseconds)"
|
||||
STASHED=1
|
||||
else
|
||||
STASHED=0
|
||||
fi
|
||||
|
||||
# Fetch and checkout
|
||||
git fetch origin "$BRANCH" 2>/dev/null || git fetch origin master
|
||||
if git rev-parse --verify "origin/$BRANCH" &>/dev/null 2>&1; then
|
||||
# Remote branch exists — fast-forward merge
|
||||
git checkout -B "$BRANCH" "origin/$BRANCH"
|
||||
elif git rev-parse --verify "$BRANCH" &>/dev/null 2>&1; then
|
||||
# Local branch or tag
|
||||
git checkout "$BRANCH"
|
||||
else
|
||||
error "Branch/tag not found: $BRANCH"
|
||||
exit 1
|
||||
fi
|
||||
ok "Checked out $BRANCH ($(git rev-parse --short HEAD))"
|
||||
|
||||
# Update submodules
|
||||
if [ -f .gitmodules ]; then
|
||||
git submodule update --init --recursive
|
||||
ok "Submodules updated"
|
||||
fi
|
||||
|
||||
# ── Build validation ──────────────────────────────────────────────────
|
||||
if [ "$NO_BUILD_CHECK" != "1" ]; then
|
||||
step "Build validation"
|
||||
info "Building nixosConfigurations.$BUILD_HOST (no link)..."
|
||||
|
||||
if nix build --no-link --print-build-logs \
|
||||
".#nixosConfigurations.${BUILD_HOST}.config.system.build.toplevel" 2>&1; then
|
||||
ok "Build succeeded for $BUILD_HOST"
|
||||
else
|
||||
error "Build failed for $BUILD_HOST"
|
||||
exit 1
|
||||
fi
|
||||
else
|
||||
warn "Build check skipped (NO_BUILD_CHECK=1)"
|
||||
fi
|
||||
|
||||
# ── Deployment ─────────────────────────────────────────────────────────
|
||||
if [ "$ACTION" = "build" ]; then
|
||||
step "Build complete (no deployment)"
|
||||
info "Use one of: switch, boot, test, dry-activate to deploy"
|
||||
exit 0
|
||||
fi
|
||||
|
||||
step "Deployment ($ACTION)"
|
||||
|
||||
# Build the nixos-rebuild command
|
||||
case "$ACTION" in
|
||||
switch|boot|test)
|
||||
nixos-rebuild "$ACTION" \
|
||||
--flake ".#$HOSTNAME" \
|
||||
--target-host "$SSH_TARGET" \
|
||||
--build-host "localhost" \
|
||||
--use-remote-sudo \
|
||||
--max-jobs 4
|
||||
;;
|
||||
dry-activate)
|
||||
nixos-rebuild dry-activate \
|
||||
--flake ".#$HOSTNAME" \
|
||||
--target-host "$SSH_TARGET" \
|
||||
--build-host "localhost" \
|
||||
--use-remote-sudo
|
||||
;;
|
||||
*)
|
||||
error "Unknown action: $ACTION"
|
||||
echo "Valid actions: switch, boot, test, build, dry-activate"
|
||||
exit 1
|
||||
;;
|
||||
esac
|
||||
|
||||
# ── Check result ───────────────────────────────────────────────────────
|
||||
DEPLOY_EXIT=$?
|
||||
if [ $DEPLOY_EXIT -eq 0 ]; then
|
||||
echo ""
|
||||
ok "Deployment to $HOSTNAME ($ACTION) completed successfully"
|
||||
case "$ACTION" in
|
||||
switch|test)
|
||||
info "Configuration is now active"
|
||||
;;
|
||||
boot)
|
||||
info "Configuration will activate on next reboot"
|
||||
;;
|
||||
dry-activate)
|
||||
info "Dry-run complete — no changes applied"
|
||||
;;
|
||||
esac
|
||||
else
|
||||
error "Deployment failed with exit code $DEPLOY_EXIT"
|
||||
exit $DEPLOY_EXIT
|
||||
fi
|
||||
|
||||
echo ""
|
||||
echo "╔══════════════════════════════════════════════╗"
|
||||
echo "║ Deployment Complete ║"
|
||||
echo "╚══════════════════════════════════════════════╝"
|
||||
info "Host: $HOSTNAME"
|
||||
info "Branch: $BRANCH ($(git rev-parse --short HEAD))"
|
||||
info "Action: $ACTION"
|
||||
info "Time: $(date -Iseconds)"
|
||||
Reference in New Issue
Block a user