Split into dedicated hardware poller + read-only consumers

Rearchitect so a single poller process is the sole owner of all SAS/SMART/
SES hardware I/O, serialized behind one gate and paced, writing results to
Redis. The API/web and MQTT publisher become pure Redis readers — they no
longer issue any subprocess and can restart freely without touching the bus.

This addresses backplane/expander stress from concurrent + restart-triggered
SMART/SES storms (the prior model re-ran a hardware sweep on every container
start, and polled sg_ses 0x02+0x07 every 60s; 0x07 errored on the IOM6/
Xyratex expanders).

- poller.py: paced sweep (inventory, SMART per-drive w/ gap, SES 0x02, ZFS,
  host/MegaRAID), startup jitter, single-instance Redis lock, LED-queue worker
- services/hwgate.py: global serialization semaphore (POLL_CONCURRENCY=1)
- services/store.py: Redis as the only producer<->consumer interface + LED queue
- services/health.py: shared drive-health classifier (fixes overview double-count)
- gate smartctl/sg_ses/zpool/ledctl; drop sg_ses 0x07 from the hot path
- routers + temps read-only from the store; main.py drops the in-process poller
- compose: 3 services (privileged poller + unprivileged app + redis) w/ pacing knobs
This commit is contained in:
2026-06-16 15:12:34 +00:00
parent 7a50d1261f
commit 0f829e0380
19 changed files with 717 additions and 464 deletions

View File

@@ -1,43 +1,59 @@
services:
jbod-monitor:
# Producer: the ONLY service that touches hardware (smartctl/sg_ses/zpool/
# ledctl). Serialized + paced; writes everything to Redis.
poller:
build: .
image: docker.adamksmith.xyz/jbod-monitor:latest
container_name: jbod-monitor
container_name: jbod-poller
restart: unless-stopped
privileged: true
pid: host
network_mode: host
command: ["python", "-u", "poller.py"]
volumes:
- /dev:/dev
- /sys:/sys:ro
- /run/udev:/run/udev:ro
# OpenBao RO token (claude-read) mounted read-only; never bake into image.
- /etc/jbod-monitor/openbao-token:/run/secrets/openbao-token:ro
environment:
- TZ=America/Denver
- UVICORN_LOG_LEVEL=info
- ZFS_USE_NSENTER=true
- REDIS_HOST=127.0.0.1
- REDIS_PORT=6379
- REDIS_DB=0
- SMART_CACHE_TTL=120
- SMART_POLL_INTERVAL=90
# Home Assistant MQTT publishing (opt-in: leave MQTT_HOST unset to disable).
# EMQX at 10.5.30.3 is reachable by IP and bridged to the Mosquitto HA
# add-on, so discovery reaches Home Assistant either way.
# Pacing — keep the SAS bus quiet. POLL_CONCURRENCY=1 fully serializes.
- POLL_SWEEP_INTERVAL=300
- POLL_DRIVE_GAP=0.5
- POLL_CONCURRENCY=1
- POLL_STARTUP_JITTER=10
- POLL_DATA_TTL=900
depends_on:
- redis
# Consumer: REST API + web UI + Home Assistant MQTT. Reads Redis only —
# unprivileged, no /dev, can restart freely without touching hardware.
app:
build: .
image: docker.adamksmith.xyz/jbod-monitor:latest
container_name: jbod-monitor
restart: unless-stopped
network_mode: host
environment:
- TZ=America/Denver
- UVICORN_LOG_LEVEL=info
- REDIS_HOST=127.0.0.1
- REDIS_PORT=6379
- REDIS_DB=0
# Home Assistant MQTT publishing (EMQX, bridged to Mosquitto HA add-on).
- MQTT_HOST=10.5.30.3
- MQTT_PORT=1883
- MQTT_DISCOVERY_PREFIX=homeassistant
- MQTT_BASE_TOPIC=jbod-monitor
- MQTT_PUBLISH_INTERVAL=60
# Broker credentials sourced from OpenBao (secret/home_assistant,
# keys: mqtt_user/mqtt_pass). Falls back to MQTT_USERNAME/MQTT_PASSWORD
# env if the read fails. Token is the read-only claude-read token.
- OPENBAO_ADDR=https://vault.adamksmith.xyz
- OPENBAO_TOKEN_FILE=/run/secrets/openbao-token
- MQTT_SECRET_PATH=home_assistant
- MQTT_SECRET_USER_KEY=mqtt_user
- MQTT_SECRET_PASS_KEY=mqtt_pass
# Broker credentials (MQTT_USERNAME/MQTT_PASSWORD). Optional: container
# starts without it (MQTT just won't authenticate).
env_file:
- path: /etc/jbod-monitor/mqtt.env
required: false
depends_on:
- redis