#!/bin/bash # ============================================================================================== # ========================== HOST2 CONFIGURATION — unRAID-Jayred365 =========================== # ============================================================================================== # HOST2-specific variables — credentials, container names, share paths, failover lists. # Sourced after master.conf — values here extend shared profile arrays and add HOST2-specific # identity, credentials, and container configuration. # # Sparse checkout (git) ensures HOST1 never receives this file. # HOST1 never sees HOST2 credentials — clean separation at the file level. # # DO NOT put shared config here — thresholds, toggles, profiles belong in master.conf. # DO NOT put HOST1 variables here — they belong in host1.conf. # # ── STATUS ──────────────────────────────────────────────────────────────────────────────────── # HOST2 is currently being rebuilt — most sections scaffolded, fill in when back online. # When ready: set FALLBACK_ENABLED=true and DAILY_RSYNC_ENABLED=true in master.conf. # # ── INDEX ───────────────────────────────────────────────────────────────────────────────────── # # ── IDENTITY & CONNECTIVITY ──────────────────────────────────────────────────────────────── # IDENTITY hostname, SSH key # EMBY container name, URL, API key # NOTIFICATIONS Discord webhook # PARTNERSHIP auth containers, backup paths # # ── RSYNC ────────────────────────────────────────────────────────────────────────────────── # DAILY SYNC SHARES media shares HOST2 owns and pushes to HOST1 # WEEKLY SYNC SHARES appdata shares synced weekly (Sunday 2:30am) # CRITICAL SYNC SHARES appdata shares synced every 30 minutes # BACKUP VERIFY shares for checksum verification against remote # HOST2 RSYNC PROFILE host2-appdata profile for HOST2-specific appdata syncs # # ── DOCKER ───────────────────────────────────────────────────────────────────────────────── # DOCKER DAILY RESTART containers restarted daily # DOCKER WEEKLY RESTART containers restarted weekly # DOCKER WATCHDOG memory limits, health URLs, required containers, ignore list # DOCKER NETWORK CONNECT networks and containers for docker_network_connect.sh # # ── FALLBACK ─────────────────────────────────────────────────────────────────────────────── # DDNS DDNS containers managed by HOST2 # INTERNET LOSS containers stopped when internet is lost # FALLBACK TIERS what HOST2 runs for HOST1 per tier # TIER DELAYS how long HOST2 must be down before each tier activates on HOST1 # RSYNC WRITEBACK HOST2 appdata synced back on handback # # ── MEDIA ────────────────────────────────────────────────────────────────────────────────── # MEDIA PERMISSIONS share list for media_shares_permissions.sh # MEDIA CLEANER folder lists for media_cleaner.sh # # ── MONITORS ─────────────────────────────────────────────────────────────────────────────── # CERTIFICATE MONITOR domains checked for SSL expiry # SMART HEALTH drives to skip in SMART monitoring # ZFS REPORT pools to exclude from ZFS health report # # ── TRANSCODES ───────────────────────────────────────────────────────────────────────────── # TRANSCODES ramdisk size, thresholds, SSD path, server array # # ── ARR STACK ────────────────────────────────────────────────────────────────────────────── # SONARR URL, API key, path map # RADARR URL, API key, path map # ARR RECOVERY per-arr recovery toggles (no Lidarr on HOST2) # # ── SYSTEM WATCHDOG ──────────────────────────────────────────────────────────────────────── # SYSTEM WATCHDOG per-host check toggles and NIC configuration # # ── RESOURCE MANAGER ─────────────────────────────────────────────────────────────────────── # RESOURCE MANAGER containers paused/stopped under memory pressure # # ============================================================================================== # ============================================================================================== # ── IDENTITY & CONNECTIVITY ─────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ Identity ━━━ # HOST2 hostname lives in master.conf (not a credential — safe for all servers). # SSH key used for all server-to-server operations — rsync, failover container commands. # Must be in /root/.ssh/ and authorised in HOST1's /root/.ssh/authorized_keys. HOST2_SSH_KEY="/root/.ssh/Jayred365-rsync-key" HOST2_OWNER="jayred365" HOST2_OWNER_EMAIL="" # fill in when HOST2 is back online # ━━━ Unraid API ━━━ # Generate in Unraid: Settings → Management Access → API Keys → + New Key HOST2_UNRAID_API_KEY="2bdf5119d61eefa3023434748bd1c171bd23dc0b2ebc8586e24abe07df986acc" # ━━━ Emby ━━━ # Referenced by transcode_manager.sh, emby_session_report.sh, emby_database_repair.sh, # weekly_sync_maintenance.sh, and HOST2_TRANSCODE_SERVERS below. # API key: Emby Dashboard → API Keys → + New Key HOST2_EMBY_CONTAINER="Emby-Jayred365" HOST2_EMBY_URL="http://localhost:8096" HOST2_EMBY_API_KEY="your-host2-emby-api-key" # ━━━ Jellyfin ━━━ # API key: Jellyfin Dashboard → Administration → API Keys → + New Key HOST2_JELLYFIN_CONTAINER="Jellyfin" HOST2_JELLYFIN_URL="http://localhost:8095" HOST2_JELLYFIN_API_KEY="956d0168987f4e4680626653abb080f0" # ━━━ Notifications ━━━ # Discord webhook — leave blank to disable. # Per-host so HOST1 and HOST2 can post to different channels or only one server notifies. HOST2_DISCORD_WEBHOOK="" # ━━━ Partnership ━━━ # HOST2 is the mirror — HOST1 is always the owner unless --transfer has been run. # See README-Partnership.md and master.conf PARTNERSHIP section for full lifecycle docs. # Auth containers reconfigured on onboard/offboard. # Format: "ContainerName|WebUIPort" # On onboard → WebUI pointed at owner's Tailscale IP (mirror clicks NPM, gets owner's auth) # On offboard → WebUI pointed back at localhost HOST2_PARTNERSHIP_AUTH_WEBUIS=( # fill in when HOST2 is back online # "NginxProxyManager|81" ) # Containers to stop on this server before the owner deploys the auth stack during onboard. # List whatever auth/proxy containers are currently running here. HOST2_PARTNERSHIP_REPLACE_CONTAINERS=( "NginxProxyManager" "Authelia" "Authelia-Secondary" "Mariadb-Authelia" "Mariadb-Authelia-Secondary" "Redis-Authelia" "Redis-Authelia-Secondary" "Lldap" ) # Arr containers to stop on this server before the owner deploys the arr stack during onboard. HOST2_PARTNERSHIP_ARR_REPLACE_CONTAINERS=( # "Sonarr" # "Radarr" # "Lidarr" # "Prowlarr" # "Bazarr" ) # Paths HOST1 should collect during the grace window after offboard. # Notified on offboard — no auto-deletion, HOST1 must collect manually within PARTNERSHIP_GRACE_HOURS. HOST2_PARTNERSHIP_MIRROR_BACKUPS=( # fill in when HOST2 is back online ) # Containers parked on this server when partnership is active. # Stopped on onboard (owner deploys its stack instead), restarted on offboard. HOST2_PARTNERSHIP_OWN_CONTAINERS=( # "Emby" # "NginxProxyManager" ) # This server's desired Emby admin account on the shared Emby instance. # Set these — owner reads them during --onboard to create the account. HOST2_PARTNERSHIP_EMBY_ADMIN_USER="" # desired Emby username HOST2_PARTNERSHIP_EMBY_ADMIN_PASS="" # desired Emby password # ============================================================================================== # ── RSYNC ───────────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ Daily Sync Shares ━━━ # Shares HOST2 pushes to all other nodes every night (1am via daily_sync_maintenance.sh). # Mesh model: every node pushes every media share — no ownership, no mirrors. # arr_sync ensures all arr libraries converge (union). rsync spreads files (additive, no --delete). # arr_cleanup removes true orphans based on local arr state. # Any node can download content to any share — it propagates to all nodes on the next cycle. # Nextcloud excluded — personal data, not arr-managed, synced HOST1→HOST2 only as offsite backup. # Uses DEFAULT_RSYNC_OPTS from master.conf — no profile needed. # For shares needing container stops or custom options — add a profile in master.conf. HOST2_DAILY_SYNC_SHARES=( /mnt/user/Anime_Movies /mnt/user/Anime_Shows /mnt/user/Books /mnt/user/Intros /mnt/user/Kids_Movies /mnt/user/Kids_Tv_Shows /mnt/user/Movies /mnt/user/Music /mnt/user/Music_Videos /mnt/user/stand-up_comedy /mnt/user/Sports /mnt/user/Tv_Shows /mnt/user/Anime_Shows-Old /mnt/user/Anime_Movies-Old ) # Personal encrypted shares — synced for offsite backup, independent of media shares. # ZFS encrypted at dataset level — remote receives encrypted blocks, cannot read content. # See README-Rsync_Setup.md for ZFS encryption setup before uncommenting. HOST2_PERSONAL_SHARES=( # /mnt/user/HOST2-Personal # uncomment after creating encrypted dataset ) # ━━━ Weekly Sync Shares ━━━ # Appdata shares synced during the weekly maintenance window (Sunday 2:30am). # Containers stopped both sides before sync — full clean state guaranteed. # Profiles drive container stops, excludes, and options — configured in master.conf RSYNC section. HOST2_WEEKLY_SYNC_SHARES=( # fill in when HOST2 is back online # "/mnt/user/Media_Server/Emby" # "/mnt/user/appdata-Fallback/Critical-Data" ) # ━━━ Intermediate Sync Shares ━━━ # Shares synced every 4 hours by intermediate_sync_maintenance.sh. # Uses DEFAULT_RSYNC_OPTS (no --delete) — for sub-daily propagation of metadata or watch state. # Full media share sync stays in the daily window. Leave empty to skip mid-day rsync entirely. HOST2_INTERMEDIATE_SYNC_SHARES=( # fill in when HOST2 is back online # Example: "/mnt/user/Emby_Metadata" ) # ━━━ Critical Sync Shares ━━━ # Appdata shares synced every 30 minutes by critical_sync_maintenance.sh. # Format: "/path/to/share" or "/path/to/share|profile-name" HOST2_CRITICAL_SYNC_SHARES=( # fill in when HOST2 is back online # "/mnt/user/appdata-Fallback/Critical-Data|critical-fallback" # "/mnt/user/Media_Server/Emby|emby-fallback" ) # ━━━ Backup Verify ━━━ # Shares verified by backup_verify.sh — random file checksum comparison against remote. # Leave empty to use HOST2_DAILY_SYNC_SHARES automatically. # Sample size and minimum file size defined in master.conf. HOST2_BACKUP_VERIFY_SHARES=( # leave empty to use HOST2_DAILY_SYNC_SHARES automatically ) # ━━━ HOST2 Rsync Profile — host2-appdata ━━━ # HOST2-specific appdata sync profile — extends the shared PROFILE_* arrays in master.conf. # Use for appdata unique to HOST2. # Shared appdata (auth stack, Emby) use dedicated profiles defined in master.conf. # Run manually: bash Rsync/rsync.sh /mnt/user/appdata-Fallback/HOST2-Appdata --profile=host2-appdata PROFILE_RSYNC_OPTS[host2-appdata]="-av --info=progress2 --bwlimit=${PROFILE_BW_LIMIT[host2-appdata]:-8000}" PROFILE_BW_LIMIT[host2-appdata]=8000 PROFILE_RETRY_COUNT[host2-appdata]=3 PROFILE_SLEEP[host2-appdata]=300 PROFILE_CRITICAL_CONTAINER_NAMES[host2-appdata]="" # fill in when HOST2 is back online PROFILE_DELAYED_CONTAINERS[host2-appdata]="" PROFILE_CONTAINER_DELAY[host2-appdata]=5 PROFILE_EXCLUDE_DIRS[host2-appdata]="logs *.tmp" # ============================================================================================== # ── DOCKER ──────────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ Docker Daily Restart ━━━ # Containers restarted every day via DAILY_MAINTENANCE_SCRIPTS. # Fill in when HOST2 is back online — add containers that degrade without daily restart. HOST2_DAILY_RESTART_CONTAINERS=( "NginxProxyManager" # add HOST2 daily restart containers here ) # ━━━ Docker Weekly Restart ━━━ # Less critical services restarted weekly via WEEKLY_MAINTENANCE_SCRIPTS (Sunday 2:30am). # Containers already stopped for weekly sync — restart adds zero extra downtime. HOST2_WEEKLY_RESTART_CONTAINERS=( # add HOST2 weekly restart containers here ) # ━━━ Docker Watchdog ━━━ # Per-HOST2 container configuration for docker_watchdog.sh. # Shared thresholds and toggles live in master.conf. # Memory hard limits in MB — immediate restart if exceeded. # Set at "container is clearly broken" not "container is busy". # 20GB=20480 16GB=16384 12GB=12288 10GB=10240 8GB=8192 4GB=4096 2GB=2048 1GB=1024 declare -A HOST2_WATCHDOG_CONTAINERS=( ["Emby"]=16384 # fill in correct limit when HOST2 is back online ) # HTTP health check URLs — checked every cycle, strike system before restart. # Only add containers with a meaningful web interface to check. declare -A HOST2_WATCHDOG_CONTAINER_URLS=( ["Emby"]="http://localhost:8096" ) # Required containers — must always be running on HOST2. # Strike system before restart — repeated failures go on skip list, auto-clears on recovery. # Listed in dependency order — dependencies before dependents. HOST2_WATCHDOG_REQUIRED_CONTAINERS=( "NginxProxyManager" # add HOST2 required containers here when back online ) # Containers to skip in Tier 2 global scan — legitimately stopped or frequently restarting. # Watchdog leaves these alone entirely — no restart attempts, no crash loop tracking. HOST2_WATCHDOG_SCAN_IGNORE=( # add HOST2 scan ignore containers here when back online ) # Dependency ordering — skip restarting a container if its dependency is also down. # Prevents watchdog from restarting dependent services before their dependencies are up. # SPACE-SEPARATED STRINGS — converted to array at runtime. declare -A HOST2_WATCHDOG_DEPENDENCIES=( # add HOST2 dependencies here when containers are defined # ["Authelia"]="Mariadb-Authelia Redis-Authelia" ) # Per-container appdata growth suppress ceilings in MB. # ONLY needed in specific cases — growth rate detection covers all containers automatically. # Use when a container legitimately has large stable data and you want to suppress false-positive # growth alerts. Add entries here only when a container triggers warnings it shouldn't. declare -A HOST2_WATCHDOG_APPDATA_SIZES=( # add HOST2 suppress entries here only as needed ) # ━━━ Docker Network Connect ━━━ # Containers connected to custom networks at array start by docker_network_connect.sh. # Networks created if they don't exist — idempotent, safe to re-run. HOST2_NETWORK_CONNECT_CONTAINERS=( # fill in when HOST2 is back online ) HOST2_NETWORK_CONNECT_NETWORKS=( # fill in when HOST2 is back online ) # ============================================================================================== # ── FALLBACK ────────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ DDNS ━━━ # DDNS containers HOST2 manages — started/stopped by fallback.sh per DDNS absolute rules: # Internet loss → stop immediately # Failover → HOST1 starts HOST2's DDNS as Tier 1 (before any other containers) # Handback → stop HOST2's DDNS on HOST1 → rsync → start containers → start local DDNS last HOST2_DDNS_CONTAINERS=( "Gmer4Lfe.us" ) # ━━━ Internet Loss ━━━ # Containers stopped immediately on HOST2 when internet connection is lost. # Prevents external-facing services from operating without connectivity. FALLBACK_HOST2_STOP_ON_NO_NET=( "Gmer4Lfe.us" ) # ━━━ Fallback Tiers — HOST2 Runs for HOST1 ━━━ # Containers HOST2 starts when HOST1 goes down. # Tier 1 is always immediate — vital services cannot wait. # Higher tiers activate after HOST1_TIER*_DELAY minutes (set in host1.conf). FALLBACK_HOST2_COVERS_HOST1_TIER1=( "Gmer4Lfe.com" "Gitea" # source of truth — must be reachable even when HOST1 auth stack is down "Emby" "VaultWarden-Gmer4Lfe" "Dispatcharr" "Dispatcharr-Basic" "Dispatcharr-Iptv-Users" "ErsatzTV-Emby" ) FALLBACK_HOST2_COVERS_HOST1_TIER2=( "Postgres-NextCloud" "NextCloud" "PostgreSQL_Immich" "Immich-Gmer4Lfe" ) FALLBACK_HOST2_COVERS_HOST1_TIER3=( "Gitea" ) FALLBACK_HOST2_COVERS_HOST1_TIER4=( "Sonarr" "Radarr" "Lidarr" "Readarr" "Prowlarr" "Bazarr" "SABnzbd-Gmer4Lfe" "Qbittorrent-Gmer4Lfe" "LidaTube" "Pinchflat" "ChannelTube" ) # ━━━ Tier Delays — HOST2's Containers on HOST1 ━━━ # How long HOST2 must be down before each tier activates on HOST1 — in minutes. # Tier 1 is always immediate — no delay var needed. HOST2_TIER2_DELAY=240 # 4 hours — productivity services HOST2_TIER3_DELAY=720 # 12 hours — secondary services HOST2_TIER4_DELAY=1440 # 24 hours — arrs + downloaders # ━━━ Rsync Writeback — HOST2 Appdata Back on Handback ━━━ # Syncs HOST2 appdata BACK to HOST2 when it comes back online after a failover. # Containers stopped before writeback — clean source, no competing writes. # # HOST2_TIER1_WRITEBACK_DELAY: short outages skip Tier 1 writeback — primary state # is more reliable than dirty sync data for brief outages. HOST2_TIER1_WRITEBACK_DELAY=60 # skip writeback if outage under 1hr # Tier 4 automatically syncs HOST2_DAILY_SYNC_SHARES — only list paths NOT in that array. FALLBACK_HOST2_WRITEBACK_TIER1=( # "/mnt/user/appdata-Fallback/Jayred365-Emby" ) FALLBACK_HOST2_WRITEBACK_TIER2=( # "/mnt/user/appdata-Fallback/Jayred365-Important" ) FALLBACK_HOST2_WRITEBACK_TIER3=( # "location-placeholder" ) FALLBACK_HOST2_WRITEBACK_TIER4=( "/mnt/user/appdata-Fallback/Arrs_Stack" # arr databases — downloads queued during outage ) # ============================================================================================== # ── MEDIA ───────────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ Media Permissions ━━━ # Shares that media_shares_permissions.sh applies PERMISSIONS_MODE and PERMISSIONS_OWNER to. # Runs first in DAILY_MAINTENANCE_SCRIPTS — arr cleanup depends on correct ownership. HOST2_MEDIA_PERMISSION_SHARES=( /mnt/user/Anime_Movies /mnt/user/Anime_Shows ) # ━━━ Media Cleaner ━━━ # Folder lists for media_cleaner.sh — two profiles: anime and media. # File patterns shared across all servers — defined in master.conf. # Called via DAILY_MAINTENANCE_SCRIPTS. Run manually: Media/media_cleaner.sh anime|media HOST2_ANIME_CLEAN_FOLDERS=( /mnt/user/Anime_Movies /mnt/user/Anime_Shows ) HOST2_MEDIA_CLEAN_FOLDERS=( # fill in when HOST2 is back online ) # ============================================================================================== # ── MONITORS ────────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # ━━━ Certificate Monitor ━━━ # Domains checked via direct openssl connection — not relying on NPM's certificate state. # Checks the actual certificate served, not what NPM thinks it has. # Thresholds (CERT_WARN_DAYS, CERT_CRIT_DAYS) defined in master.conf. HOST2_CERT_MONITOR_DOMAINS=( # fill in when HOST2 is back online ) # ━━━ SMART Health ━━━ # Drives skipped in SMART attribute monitoring — hardware is server-specific. # Thresholds read from dynamix.cfg at runtime — fallbacks in master.conf. HOST2_SMART_IGNORE_DRIVES=( "sda" # boot USB — SMART not meaningful on flash drives ) # ━━━ ZFS Report ━━━ # Pools excluded from the weekly ZFS health report — reduces noise from single-disk array pools. # Pool health thresholds defined in master.conf. HOST2_ZFS_REPORT_IGNORE_POOLS=( # fill in when HOST2 is back online ) # ============================================================================================== # ── TRANSCODES ──────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # Ramdisk size ceiling — tmpfs only uses RAM actually needed, not the full size upfront. # Adjust HOST2_RAMDISK_WARN_GB and HOST2_RAMDISK_LOW_GB together if this changes. HOST2_RAMDISK_SIZE="8G" # Usage thresholds — coupled to HOST2_RAMDISK_SIZE, adjust all three together if size changes. # Hysteresis gap (6.8 - 5.5 = 1.3GB) prevents flip-flop between ramdisk and SSD. HOST2_RAMDISK_WARN_GB=6.8 # flip to SSD when ramdisk usage reaches this HOST2_RAMDISK_LOW_GB=5.5 # flip back to ramdisk when usage drops to this # SSD fallback path — where transcodes land when ramdisk exceeds HOST2_RAMDISK_WARN_GB. # Must be on cache pool — array disks too slow for active transcode writes. HOST2_TRANSCODE_SSD="/mnt/cache/Temp_Storage/Emby/Transcodes/" # Media servers sharing the ramdisk transcode space on HOST2. # Format: "ContainerName|URL|APIKey|Type" — Type: emby | jellyfin | plex # Entries with placeholder API keys are skipped automatically. # ⚠️ Tdarr does NOT belong here — keep Tdarr on SSD, not ramdisk. HOST2_TRANSCODE_SERVERS=( "${HOST2_EMBY_CONTAINER}|${HOST2_EMBY_URL}|${HOST2_EMBY_API_KEY}|emby" "${HOST2_JELLYFIN_CONTAINER}|${HOST2_JELLYFIN_URL}|${HOST2_JELLYFIN_API_KEY}|jellyfin" ) # ============================================================================================== # ── ARR STACK ───────────────────────────────────────────────────────────────────────────────── # ============================================================================================== # Used by arr cleanup scripts and arrs_failed_stalled_recovery.sh. # detect_hosts() selects HOST2 vars when running on HOST2. # Lidarr does not run on HOST2 — HOST1_LIDARR_RECOVERY flag handles the exit cleanly. # # PATH MAPS — container path → host path translation. # Arr stores file paths using container-internal paths — scripts need host paths to scan. # Add one entry per root folder in arr Settings → Media Management → Root Folders. HOST2_FANART_API_KEY="Yd7147a43b692df0b364b94dc47efb81" HOST2_LASTFM_API_KEY="be6dc169c33ae263e690c30d18b7491d" # ━━━ Sonarr ━━━ HOST2_SONARR_URL="http://localhost:8989" HOST2_SONARR_API_KEY="130decd3db5b4c25afad64864cd03f9f" HOST2_SONARR_TV_ROOT="/mnt/user/Anime_Shows" declare -A HOST2_SONARR_PATH_MAP=( # fill in when HOST2 is back online # ["/tv"]="/mnt/user/Anime_Shows" ) # ━━━ Radarr ━━━ HOST2_RADARR_URL="http://localhost:7878" HOST2_RADARR_API_KEY="d43a3ec6cf1549edb4af0cc63f98b2a9" HOST2_RADARR_MOVIES_ROOT="/mnt/user/Anime_Movies" declare -A HOST2_RADARR_PATH_MAP=( # fill in when HOST2 is back online # ["/anime-movies"]="/mnt/user/Anime_Movies" ) # ━━━ Arr Recovery Toggles ━━━ # false = skip that arr on this host — exits cleanly without error HOST2_SONARR_RECOVERY=true HOST2_RADARR_RECOVERY=true # HOST2_LIDARR_RECOVERY not set — Lidarr does not run on HOST2 # ============================================================================================== # ── SYSTEM WATCHDOG ─────────────────────────────────────────────────────────────────────────── # ============================================================================================== # Per-host check toggles and NIC config for system_watchdog.sh. # Aliased by detect_hosts() — script uses unprefixed SYS_WATCHDOG_* names. # HOST2: i5 10th gen 64GB — being rebuilt, lighter workload, no ZFS cache pools. # # Conservative defaults during rebuild — re-enable checks as HOST2 stabilises. # Three-tier response — all critical checks enabled regardless of rebuild state: # Tier 1 (bypass strikes, reboot now): docker daemon, rootfs full, kernel oops, FD, /boot # Tier 2 (bypass strikes with OOM): RAM critical + OOM kills in cycle # Tier 3 (standard strike system): selectively disabled during rebuild # # RAM tiers, OOM limits, and reboot loop settings in master.conf System Watchdog section. # ━━━ Primary NIC ━━━ # Network interface for NIC state check — verify with: ip link show | grep "^[0-9]" # Common values: eth0, bond0, br0, eno1 HOST2_SYS_WATCHDOG_NIC="eth0" # ━━━ Tier 1 — Critical Checks ━━━ # All critical checks always enabled — these protect against acute failure regardless of # rebuild state. Disabling any is not recommended. # Docker daemon unresponsive → try restart, reboot if restart fails. HOST2_SYS_WATCHDOG_CHECK_DOCKER_DAEMON=true # rootfs at critical threshold (ROOTFS_CRITICAL_PCT=99) → reboot immediately. HOST2_SYS_WATCHDOG_CHECK_ROOTFS=true # Kernel BUG/Oops in dmesg delta since last cycle → reboot immediately. HOST2_SYS_WATCHDOG_CHECK_KERNEL_OOPS=true # File descriptor exhaustion at FD_CRITICAL_PCT (95%) → reboot immediately. HOST2_SYS_WATCHDOG_CHECK_FD=true # /boot read-only detected → reboot immediately. HOST2_SYS_WATCHDOG_CHECK_BOOT=true # ━━━ Tier 2 — Urgent OOM Check ━━━ # Both must be enabled for Tier 2 bypass to function. # Track kernel OOM kills each cycle via /proc/vmstat oom_kill delta. HOST2_SYS_WATCHDOG_CHECK_OOM=true # Free RAM check — 64GB RAM on HOST2, tiers adjusted relative to HOST1. # Update master.conf SYS_WATCHDOG_MEM_* thresholds if HOST2 needs different values. # Currently inheriting shared master.conf values — may want lower thresholds on 64GB. HOST2_SYS_WATCHDOG_CHECK_RAM=true # ━━━ Tier 3 — Standard Checks (strike system) ━━━ # Several checks disabled during rebuild — enable progressively as HOST2 stabilises. # Each check must fail SYS_WATCHDOG_STRIKE_LIMIT consecutive cycles before action. # /var/log filesystem usage above SYS_WATCHDOG_LOG_PCT. HOST2_SYS_WATCHDOG_CHECK_LOG=true # ZFS ARC memory check. # DISABLED — HOST2 has no ZFS cache pools. Enable if ZFS pools are added later. HOST2_SYS_WATCHDOG_CHECK_ARC=false # CPU temperature above SYS_WATCHDOG_CPU_TEMP_MAX (95°C). HOST2_SYS_WATCHDOG_CHECK_CPU_TEMP=true # Load average above SYS_WATCHDOG_LOAD_MULTIPLIER × core count. # DISABLED — rebuild operations cause legitimate load spikes. Enable after rebuild. HOST2_SYS_WATCHDOG_CHECK_LOAD=false # Zombie process count above SYS_WATCHDOG_ZOMBIE_LIMIT (50). HOST2_SYS_WATCHDOG_CHECK_ZOMBIES=true # docker_watchdog.sh persistent skip list check. # DISABLED during rebuild — skip list may be unreliable mid-rebuild, avoid false reboots. # Enable once HOST2 is fully operational and docker_watchdog.sh is running stably. HOST2_SYS_WATCHDOG_CHECK_CONTAINERS=false # /tmp filesystem usage with auto-clear attempt. HOST2_SYS_WATCHDOG_CHECK_TMP=true # Array disk error count delta in /proc/mdstat. HOST2_SYS_WATCHDOG_CHECK_MDSTAT=true # Primary NIC operstate — uses HOST2_SYS_WATCHDOG_NIC above. HOST2_SYS_WATCHDOG_CHECK_NETWORK=true # sshd running check — restart attempt before escalating. HOST2_SYS_WATCHDOG_CHECK_SSHD=true # Runaway process detection. # DISABLED — rebuild workloads may legitimately peg CPU. Enable after rebuild. HOST2_SYS_WATCHDOG_CHECK_RUNAWAY=false # ============================================================================================== # ── RESOURCE MANAGER ────────────────────────────────────────────────────────────────────────── # ============================================================================================== # Containers to manage under pressure — see master.conf RW_CRITICAL_CONTAINERS for exclusions. # docker pause at medium pressure (RAM < RW_RAM_MEDIUM_GB or load > medium threshold) # Suspended in-place — instant to pause/unpause, no state lost, no restart delay. HOST2_RW_PAUSE_CONTAINERS=( # fill in when HOST2 is back online ) # docker stop at hard pressure (RAM < RW_RAM_HARD_GB) # Full stop — these are optional/heavy services that free significant RAM when stopped. # resource_watchdog.sh restarts them when pressure fully clears (RAM >= RW_RAM_RECOVER_GB). HOST2_RW_STOP_CONTAINERS=( # fill in when HOST2 is back online ) # ============================================================================================== # ──────────────────────── End Of HOST2 Variables ────────────────────────────────────────────── # ==============================================================================================