updated failover to have same ddns logic as old script

This commit is contained in:
2026-04-18 10:15:45 -04:00
parent 3581396378
commit cd41f59334
2 changed files with 976 additions and 637 deletions
+661 -427
View File
File diff suppressed because it is too large Load Diff
+288 -183
View File
@@ -33,7 +33,7 @@
# ── DOCKER ESSENTIALS ──────────────────────────────────────────────────────────────────────
# DOCKER DAILY RESTART Containers restarted daily
# DOCKER WEEKLY RESTART Containers restarted weekly
# DOCKER WATCHDOG Container health monitoring — memory, CPU, HTTP
# DOCKER WATCHDOG Two-tier self-healing container monitoring
# DOCKER NETWORK CONNECT Connect containers to extra networks on boot
#
# ── UNRAID ESSENTIALS ──────────────────────────────────────────────────────────────────────
@@ -53,10 +53,11 @@
# ── TRANSCODES ─────────────────────────────────────────────────────────────────────────────
# TRANSCODE MANAGER Ramdisk and SSD fallback transcode management
#
# ── MONITORS ───────────────────────────────────────────────────────────────────────────────
# ── MONITORS ───────────────────────────────────────────────────────────────────────────────
# CERTIFICATE MONITOR SSL certificate expiry monitoring
# BACKUP VERIFY Random sample checksum verification against remote
# SMART HEALTH Drive SMART attribute monitoring
# ZFS MEMORY SNAPSHOT Weekly ZFS health and memory diagnostic report
# BANDWIDTH MONITOR Daily rsync transfer logging and weekly summary
# HEALTH DIGEST Aggregated system health digest — always/smart/weekly
# EMBY SESSION REPORT Weekly Emby usage statistics via API
@@ -290,66 +291,218 @@ declare -A PROFILE_SKIP_DISK_CHECK=(
# ==============================================================================================
# ── FAILOVER ──────────────────────────────────────────────────────────────────────────────────
# ==============================================================================================
# Mutual container failover between two unRAID servers 50 miles apart.
# Mutual container failover between two unRAID servers.
# Each server runs Failover/failover.sh independently — no coordination between servers.
# All decisions are based solely on two ping checks: remote reachable + internet reachable.
# All decisions based solely on two pings: remote reachable + internet reachable.
#
# States:
# NORMAL — remote up, internet up — own containers only, silent operation
# FAILOVER — remote down, internet up — start remote's containers locally (additive)
# NO_INTERNET — internet down — stop public-facing containers, wait for recovery
# DARK — remote down + internet down — same actions as NO_INTERNET
# ── STATES ────────────────────────────────────────────────────────────────────────────────────
# NORMAL — remote up, internet up — own containers only, DDNS ON, silent
# FAILOVER — remote down, internet up — start remote containers (tiered by time)
# NO_INTERNET — internet down — stop own DDNS immediately, wait for recovery
# DARK — remote down AND internet down — same as NO_INTERNET
#
# Handback: strike confirmation → pre-flight → rsync → start remote → stop local
# ── DDNS RULES — ABSOLUTE ─────────────────────────────────────────────────────────────────────
# Each server owns its own DDNS — ON when that server has internet
# Script controls DDNS exclusively — network state NEVER auto-starts DDNS
# Internet loss → stop own DDNS immediately
# Failover → start remote DDNS as first action (Tier 1)
# Handback → stop remote DDNS FIRST → rsync → start local containers
# → start local DDNS LAST — only after containers confirmed up
# One DDNS per domain active at all times — never two, never zero for long
# 1 minute TTL + 1 minute check interval = minimal user impact
#
# ── HANDBACK SEQUENCE ─────────────────────────────────────────────────────────────────────────
# Strike confirmation → pre-flight → stop remote DDNS → stop remote containers
# → rsync writeback → start local containers → start local DDNS → NORMAL
# Containers only down during rsync window — minimise this time
#
# ── TIERED FAILOVER ───────────────────────────────────────────────────────────────────────────
# Tier 1 — Immediate — vital services + Live TV — people are watching, can't wait
# Tier 2 — configurable delay — shared productivity services
# Tier 3 — configurable delay — secondary services
# Tier 4 — configurable delay — arrs + downloaders — workflow continuity
# Delays set independently per host below
EXTERNAL_IP="8.8.8.8" # external IP to ping for internet connectivity check
FAILOVER_CHECK_INTERVAL=120 # seconds between state checks
FAILOVER_HANDBACK_STRIKES=2 # consecutive remote-up confirmations before handback
# 1 minute TTL + 2 minute interval = minimal gap
FAILOVER_HANDBACK_STRIKES=2 # consecutive remote-up checks before handback
# 2 strikes x 120s = 4 min confirmation window
FAILOVER_STATE_FILE="/boot/config/failover_state.db"
# persists on /boot/ — survives reboots
# tracks: state, failover_start, strikes, tier flags
# ━━━ Failover Test ━━━
# Used by Failover/failover_test.sh — controlled simulation of the failover lifecycle.
# failover_test.sh blocks remote connectivity via iptables then observes failover.sh behavior.
# Used by Failover/failover_test.sh — controlled simulation via iptables block.
# All failover logic stays in failover.sh — test script is the harness only.
#
# ⚠️ Run during a maintenance window — real containers start and stop during the test.
# ⚠️ Run during maintenance window — real containers start and stop during the test.
# Use --dry-run first to walk through phases without touching anything.
FAILOVER_TEST_BLOCK_WAIT=150 # seconds to hold iptables block
# must be > FAILOVER_CHECK_INTERVAL + buffer
FAILOVER_TEST_HANDBACK_WAIT=360 # seconds to wait for handback completion
# covers FAILOVER_HANDBACK_STRIKES x INTERVAL + rsync
# Seconds to hold the iptables block — must be longer than FAILOVER_CHECK_INTERVAL
# so failover.sh has time to detect the outage and change state
FAILOVER_TEST_BLOCK_WAIT=150 # 150s = FAILOVER_CHECK_INTERVAL + 30s buffer
# ━━━ DDNS — Script Controlled Exclusively ━━━
# Each server owns its own DDNS containers — one domain per server.
# DDNS is started and stopped ONLY by this script — never by network state returning.
# HOST1 DDNS starts last in handback (after containers confirmed up).
# HOST1 DDNS stops first on internet loss.
# HOST2 DDNS starts when HOST2 detects HOST1 is down (Tier 1).
# HOST2 DDNS stops before handback rsync begins.
# Seconds to wait for handback after restoring connectivity
# Must cover FAILOVER_HANDBACK_STRIKES x FAILOVER_CHECK_INTERVAL plus rsync time
# 2 strikes x 120s = 240s minimum — add buffer for rsync handback jobs
FAILOVER_TEST_HANDBACK_WAIT=360 # 360s = 6 minutes — adjust if rsync takes longer
FAILOVER_HOST1_STARTS_FOR_HOST2=(
"Gmer4Lfe.com"
"Gmer4Lfe.us"
HOST1_DDNS_CONTAINERS=(
"Gmer4Lfe.com" # HOST1's own DDNS — ON when HOST1 has internet
# covers gmer4lfe.com pointing to HOST1 IP
)
HOST2_DDNS_CONTAINERS=(
"Gmer4Lfe.us" # HOST2's own DDNS — ON when HOST2 has internet
# covers gmer4lfe.us pointing to HOST2 IP
)
# ━━━ Containers to stop on internet loss ━━━
# Own DDNS handled separately above — list additional containers here if needed
# These stop when this server loses internet — regardless of remote state
FAILOVER_HOST1_STOP_ON_NO_NET=(
"Gmer4Lfe.com"
"Gmer4Lfe.us"
)
FAILOVER_HOST1_RSYNC_JOBS=(
# "/mnt/user/appdata-Failover/Jayred365"
# "/mnt/user/Media_Server/Emby-Jayred"
# "container-name" # add containers that should stop without internet
)
# HOST2 (unRAID-Jayred365 — Secondary)
FAILOVER_HOST2_STARTS_FOR_HOST1=(
"Emby"
"Gmer4Lfe.com"
"Gmer4Lfe.us"
)
FAILOVER_HOST2_STOP_ON_NO_NET=(
"Gmer4Lfe.com"
"Gmer4Lfe.us"
# "container-name"
)
FAILOVER_HOST2_RSYNC_JOBS=(
# "/mnt/user/appdata-Failover/Gmer4Lfe"
# ━━━ HOST1 runs these for HOST2 when HOST2 goes down ━━━
# HOST2's DDNS listed in Tier 1 — starts immediately as first action
# List HOST2's specific services here — HOST2's own containers only
# Do NOT list shared services that HOST1 already runs
# Tier 1 — Immediate — starts as soon as HOST2 is detected down
FAILOVER_HOST1_RUNS_FOR_HOST2_IMMEDIATE=(
"Gmer4Lfe.us" # HOST2's DDNS — start first, covers HOST2's domain
"VaultWarden-Jayred365" # HOST2's password manager — immediate access needed
# "container-placeholder" # add HOST2 specific services here
)
# Tier 2 — starts after HOST2_TIER2_DELAY minutes
FAILOVER_HOST1_RUNS_FOR_HOST2_2HR=(
# "container-placeholder"
)
# Tier 3 — starts after HOST2_TIER3_DELAY minutes
FAILOVER_HOST1_RUNS_FOR_HOST2_6HR=(
# "container-placeholder"
)
# Tier 4 — starts after HOST2_TIER4_DELAY minutes
FAILOVER_HOST1_RUNS_FOR_HOST2_18HR=(
# "container-placeholder"
)
# ━━━ HOST2 runs these for HOST1 when HOST1 goes down ━━━
# HOST1's DDNS listed in Tier 1 — starts immediately to cover HOST1's domain
# Live TV in Tier 1 — people are watching, cannot wait for tiered startup
# Auth stack in Tier 1 — everything proxied through NPM needs auth
# Tier 1 — Immediate — vital services and Live TV cannot wait
FAILOVER_HOST2_RUNS_FOR_HOST1_IMMEDIATE=(
"Gmer4Lfe.com" # HOST1's DDNS — start first, covers HOST1's domain
"Emby" # media server — users are actively watching
"NginxProxyManager" # reverse proxy — all services route through this
"Lldap-Gmer4Lfe" # auth directory — required by Authelia
"Mariadb-Authelia" # auth database — required by Authelia
"Redis-Authelia" # auth cache — required by Authelia
"Authelia" # authentication — required for all proxied services
"Authelia-Secondary" # auth redundancy
"Redis-Authelia-Secondary" # auth secondary cache
"VaultWarden-Gmer4Lfe" # password manager — critical, immediate access needed
"Dispatcharr" # Live TV — people are watching, cannot wait
"Dispatcharr-Basic" # Live TV basic profile
"Dispatcharr-Iptv-Users" # Live TV IPTV users
"ErsatzTV-Emby" # Live TV scheduling and channel management
)
# Tier 2 — starts after HOST1_TIER2_DELAY minutes
# Productivity services — important but can wait a couple of hours
FAILOVER_HOST2_RUNS_FOR_HOST1_2HR=(
"Postgres-NextCloud" # NextCloud database — must start before NextCloud
"NextCloud" # file access and collaboration
"PostgreSQL_Immich" # Immich database
"Immich-Gmer4Lfe" # photo management
"Jellyseerr" # media request management
# "container-placeholder"
)
# Tier 3 — starts after HOST1_TIER3_DELAY minutes
# Secondary services — useful but not immediately critical
FAILOVER_HOST2_RUNS_FOR_HOST1_6HR=(
"Organizrv2-Gmer4Lfe" # dashboard — nice to have
"AdGuard-Home" # DNS filtering
"UptimeKuma-Gmer4Lfe" # uptime monitoring
"Gitea" # git server
"Collabora-CODE" # document editing for NextCloud
# "container-placeholder"
)
# Tier 4 — starts after HOST1_TIER4_DELAY minutes
# Full workflow mode — arrs and downloaders
# Minimal writeback on handback — start fresh is cleaner than syncing download state
FAILOVER_HOST2_RUNS_FOR_HOST1_18HR=(
"Sonarr" # TV show management
"Radarr" # movie management
"Lidarr" # music management
"Readarr" # book management
"Prowlarr" # indexer management
"Bazarr" # subtitle management
"SABnzbd-Gmer4Lfe" # usenet downloader
"Qbittorrent-Gmer4Lfe" # torrent downloader
"LidaTube" # YouTube music downloader
"Pinchflat" # YouTube channel downloader
"ChannelTube" # YouTube channel management
# "container-placeholder"
)
# ━━━ Tier Delay Settings ━━━
# How long primary must be down before each tier activates — set in minutes
# Tier 1 is always immediate — no delay
# Set independently per host — a large server may want longer delays than a small one
# Adjust based on your tolerance for resource usage on the covering server
# Delays for HOST1's containers running on HOST2 (HOST1 is down)
HOST1_TIER2_DELAY=120 # 2 hours — NextCloud, Immich can wait
HOST1_TIER3_DELAY=360 # 6 hours — dashboard, monitoring, Gitea
HOST1_TIER4_DELAY=1080 # 18 hours — full workflow, arrs and downloaders
# Delays for HOST2's containers running on HOST1 (HOST2 is down)
HOST2_TIER2_DELAY=120
HOST2_TIER3_DELAY=360
HOST2_TIER4_DELAY=1080
# ━━━ Rsync Writeback Jobs ━━━
# Run during handback — syncs critical appdata back to primary before containers restart.
# Containers are stopped before this runs — clean source, no competing writes.
# Full bandwidth available — DDNS stopped, containers stopped, nothing competing.
#
# Priority:
# Critical — Emby userdata/playstates (small, fast, important)
# Critical — Auth stack data
# Skip — Media files (already on primary, never moved)
# Skip — Downloads (start fresh — cleaner than syncing partial state)
#
# Format: "/path/to/source" — matched to rsync profile by directory basename
# HOST1 writeback — run by HOST2 during HOST1 handback
FAILOVER_HOST1_WRITEBACK=(
"/mnt/user/appdata-Failover/Critical-Data" # auth stack — Authelia, Mariadb, Redis, LLDAP, NPM
"/mnt/user/appdata-Failover/Important-Data" # NextCloud + Postgres
"/mnt/user/appdata-Failover/Emby" # Emby userdata, playstates, metadata
"/mnt/user/appdata-Failover/Gmer4Lfe" # server specific appdata — Organizr, UptimeKuma
# "/mnt/user/appdata-Failover/Arrs_Stack" # skip — arrs start fresh on handback
)
# HOST2 writeback — run by HOST1 during HOST2 handback
FAILOVER_HOST2_WRITEBACK=(
# "/mnt/user/appdata-Failover/Jayred365" # HOST2 specific appdata
# "container-placeholder"
)
# ==============================================================================================
@@ -357,9 +510,8 @@ FAILOVER_HOST2_RSYNC_JOBS=(
# ==============================================================================================
# ━━━ Docker Daily Restart ━━━
# Containers restarted every day by Docker_Essentials/docker_daily_restart.sh.
# Keeps services fresh and clears memory leaks that accumulate over time.
# Case-sensitive — must match exact Docker container names shown in the unRAID Docker tab.
# Containers restarted every day — keeps services fresh, clears memory leaks.
# Case-sensitive — must match exact Docker container names in the unRAID Docker tab.
DAILY_RESTART_CONTAINERS=(
"NginxProxyManager"
"Authelia"
@@ -370,8 +522,8 @@ DAILY_RESTART_CONTAINERS=(
)
# ━━━ Docker Weekly Restart ━━━
# Containers restarted once per week by Docker_Essentials/docker_weekly_restart.sh.
# For less critical services that benefit from periodic restart but don't need daily cycling.
# Less critical services that benefit from periodic restart but don't need daily cycling.
# Case-sensitive — must match exact Docker container names in the unRAID Docker tab.
WEEKLY_RESTART_CONTAINERS=(
"NextCloud"
"Organizrv2-Gmer4Lfe"
@@ -380,20 +532,18 @@ WEEKLY_RESTART_CONTAINERS=(
)
# ━━━ Docker Watchdog ━━━
# Two-tier self-healing container monitoring:
# Tier 1 — strict monitoring of explicitly configured containers
# Tier 2 — global health scan of ALL running containers
# Two-tier self-healing container monitoring.
# Tier 1 — strict monitoring of explicitly configured containers
# Tier 2 — global health scan of ALL running containers
#
# Cross-cutting intelligence applies to both tiers:
# Cross-cutting intelligence:
# Startup grace — skip restarts while system is still booting
# Dependency order — restart database before app, not the other way around
# Dependency order — restart database before app
# Restart loop — stop restarting after limit hit → skip list → notify critical
# Skip list — persistent across reboots, auto-clears when container recovers
# Batch notify — one clean summary per run instead of one ping per event
# Batch notify — one clean summary per run
# ── Tier 1 — Strict Monitoring ────────────────────────────────────────────────────────────
# Memory hard limits in MB — immediate restart if exceeded, no strike system
# Memory hard limits in MB — immediate restart if exceeded
# 20GB=20480 16GB=16384 14GB=14336 12GB=12288 10GB=10240
# 8GB=8192 6GB=6144 4GB=4096 2GB=2048 1GB=1024
declare -A WATCHDOG_CONTAINERS=(
@@ -403,13 +553,13 @@ declare -A WATCHDOG_CONTAINERS=(
["Code-Server"]=1024
)
# HTTP responsiveness checks — omit container to skip its HTTP check
declare -A WATCHDOG_CONTAINER_URLS=(
["Emby"]="http://localhost:8096"
)
# Containers that must always be running — strike system, persistent skip list on /boot/
# Skip list auto-clears when container recovers — no manual intervention for normal recovery
# These are your core auth and proxy stack — everything depends on them being up
WATCHDOG_REQUIRED_CONTAINERS=(
"NginxProxyManager"
"Lldap-Gmer4Lfe"
@@ -420,22 +570,27 @@ WATCHDOG_REQUIRED_CONTAINERS=(
"Redis-Authelia-Secondary"
)
# Strike thresholds
# Strike state file — /tmp resets on reboot which is correct for strike tracking
WATCHDOG_STATE_FILE="/tmp/container_watchdog_state.db"
SOFT_CPU_THRESHOLD=80 # warn at this % of total system CPU
HARD_CPU_THRESHOLD=85 # strike at this % of total system CPU
CPU_FAIL_LIMIT=2 # consecutive CPU strikes before restart
SOFT_MEM_THRESHOLD=80 # warn when container reaches this % of hard limit
RESP_FAIL_LIMIT=2 # consecutive failed HTTP checks before restart
CURL_TIMEOUT=5 # seconds before curl gives up per check
# ── Tier 2 — Global Health Scan ───────────────────────────────────────────────────────────
# CPU thresholds — normalised against total core count automatically at runtime
SOFT_CPU_THRESHOLD=80 # warn at this % of total system CPU
HARD_CPU_THRESHOLD=85 # strike at this % of total system CPU
CPU_FAIL_LIMIT=2 # consecutive CPU strikes before container restart
# Master toggle — false disables Tier 2 entirely
# Memory soft threshold — warn when container reaches this % of its hard limit
# Hard limit exceeded triggers immediate restart regardless of strikes
SOFT_MEM_THRESHOLD=80
# HTTP responsiveness check settings
RESP_FAIL_LIMIT=2 # consecutive failed curl checks before restart
CURL_TIMEOUT=5 # seconds before curl gives up per check
# Tier 2 master toggle — false disables global scan entirely
WATCHDOG_SCAN_ALL=true
# Containers to skip in Tier 2 — add containers expected to be in a non-running state
# or managed by other systems that should not be auto-restarted
# Containers to skip in Tier 2 scan entirely
# Add intentionally stopped containers or containers managed by other systems
WATCHDOG_SCAN_IGNORE=(
# "container-name"
)
@@ -450,23 +605,21 @@ WATCHDOG_SCAN_IGNORE=(
# Crash loop threshold — notify critical if Docker has restarted container this many times
WATCHDOG_CRASH_LIMIT=5
# ── Cross-cutting Intelligence ────────────────────────────────────────────────────────────
# Startup grace period — skip restarts while system is still booting
# Startup grace — skip restarts while system is still booting
# Prevents false positives while containers are coming up after array start
WATCHDOG_STARTUP_GRACE=300 # seconds after boot before watchdog acts on failures
WATCHDOG_STARTUP_GRACE=600 # seconds after boot before watchdog acts on failures
# Restart loop protection — stops hammering broken containers
# Tracks watchdog-initiated restarts per container in a bounded /boot/ file
# After limit hit → container added to skip list → notify critical → manual intervention
# After limit hit → skip list → notify critical → manual intervention needed
# Skip list auto-clears when container is found running again
WATCHDOG_CONTAINER_RESTART_LIMIT=3 # max watchdog restarts allowed in window
WATCHDOG_CONTAINER_RESTART_WINDOW=1 # hours — rolling window for restart count
WATCHDOG_CONTAINER_RESTART_LOG="/boot/config/container_restart_history.db"
# /boot/ survives reboots — bounded, auto-purges old entries
# /boot/ survives reboots — bounded, auto-purges
# Dependency ordering — skip restarting a container if its dependency is also down
# Dependency gets restarted first, dependent picked up on the next watchdog cycle
# Prevents Authelia restarting before its database is ready — it would just fail again
# Format: ["dependent"]="dependency1 dependency2"
declare -A WATCHDOG_DEPENDENCIES=(
["Authelia"]="Mariadb-Authelia Redis-Authelia"
@@ -480,20 +633,10 @@ declare -A WATCHDOG_DEPENDENCIES=(
WATCHDOG_BATCH_NOTIFY=true
# ━━━ Docker Network Connect ━━━
NETWORK_CONNECT_CONTAINERS=(
"memcached"
"Npm-CrowdSec"
)
NETWORK_CONNECT_NETWORKS=(
"nextcloud-aio"
)
# ━━━ Docker Network Connect ━━━
# Connects containers to extra Docker networks on array start.
# Connects containers to extra Docker networks on array start — many-to-many.
# Every container in the list connects to every network in the list.
# Useful when containers need to communicate across networks they were not originally
# configured with — e.g. memcached needing access to the nextcloud-aio network.
# Every container in the list connects to every network in the list (many-to-many).
# Comment out entries to disable without removing them.
NETWORK_CONNECT_CONTAINERS=(
"memcached"
"Npm-CrowdSec"
@@ -508,42 +651,37 @@ NETWORK_CONNECT_NETWORKS=(
# ==============================================================================================
# ━━━ Reboot ━━━
# Seconds of warning broadcast to all logged-in users before server_reboot.sh reboots.
# Gives users time to save work or finish what they are doing before the system goes down.
# Seconds of warning broadcast to logged-in users before server_reboot.sh reboots.
# Gives users time to save work before the system goes down.
REBOOT_SLEEP=300
# ━━━ Mover ━━━
# Seconds to wait after warning users before mover_stop.sh sends SIGTERM to the mover.
# Seconds to wait before mover_stop.sh sends SIGTERM to the mover process.
# Gives the mover time to finish its current file operation cleanly before being killed.
MOVER_STOP_TIMEOUT=300
# ━━━ Syslog Filter ━━━
# Path for the rsyslog filter file created by docker_syslog_filter.sh.
# The filter suppresses noisy Docker veth and docker0 messages from syslog on boot.
# Without this filter, every Docker network interface change floods the syslog.
# Path for the rsyslog filter file that suppresses Docker veth noise from syslog.
# Without this filter every Docker network interface change floods the syslog on boot.
FILTER_FILE="/etc/rsyslog.d/ignore-docker-veth.conf"
# ━━━ PHP-FPM ━━━
# Config file path and max children value for php_fpm_max_children.sh.
# Higher max_children allows more concurrent PHP requests to the unRAID WebGUI.
# Set based on available RAM — too high can cause memory pressure on low-RAM systems.
PHP_CONF="/etc/php-fpm.d/www.conf"
PHP_MAX_CHILDREN=250
# ━━━ Clear Logs ━━━
# System log files cleared by clear_logs.sh — Docker container logs are cleared too.
# Run weekly to prevent logs from filling the rootfs over time.
# System log files cleared weekly to prevent rootfs fill over time.
LOG_FILES=(/var/log/syslog /var/log/messages /var/log/dmesg)
# ━━━ WebGUI Watchdog ━━━
# Monitors the unRAID WebGUI and restarts services if it becomes unresponsive.
# Escalation path: nginx restart → recheck → emhttp restart → recheck → notify warning.
# emhttp is the core unRAID management daemon — restarting it is more disruptive than nginx
# but both recover cleanly. Notification sent on any restart so you know what happened.
WEBGUI_URL="http://localhost" # adjust port if non-standard e.g. http://localhost:8080
# Escalation: nginx restart → recheck → emhttp restart → recheck → notify warning.
# emhttp is the core unRAID daemon — restarting is more disruptive but recovers cleanly.
WEBGUI_URL="http://localhost" # adjust if running non-standard port
WEBGUI_TIMEOUT=5 # seconds before curl gives up on the WebGUI check
WEBGUI_NGINX_WAIT=15 # seconds to wait after nginx restart before rechecking
WEBGUI_EMHTTP_WAIT=30 # seconds to wait after emhttp restart — emhttp takes longer
WEBGUI_EMHTTP_WAIT=30 # seconds to wait after emhttp restart — takes longer
# ==============================================================================================
# ── MEDIA ─────────────────────────────────────────────────────────────────────────────────────
@@ -553,11 +691,11 @@ NETWORK_CONNECT_NETWORKS=(
# Mode and owner applied recursively to all shares in MEDIA_PERMISSION_SHARES.
# Run by Media/media_shares_permissions.sh via the media_management.sh orchestrator.
# 777 and nobody:users is standard for unRAID media shares accessible by Docker containers.
# Applied recursively so large shares take time — run overnight via orchestrator.
PERMISSIONS_MODE="777"
PERMISSIONS_OWNER="nobody:users"
# Shares to apply permissions to — add or remove paths as your library grows.
# These are applied recursively so large shares take time — run overnight via orchestrator.
MEDIA_PERMISSION_SHARES=(
/mnt/user/Anime_Movies
/mnt/user/Anime_Movies-Old
@@ -616,7 +754,8 @@ ANIME_FILE_PATTERNS=(
'*.log' '*.json'
)
# File patterns deleted by the media profile — includes *.iso and *.lrc not needed in anime
# File patterns deleted by the media profile
# Includes *.iso and *.lrc not needed in anime profile
MEDIA_FILE_PATTERNS=(
'*.sfv' '*.md5' '*.sha1' '*.txt' '*.url' '*.lnk'
'*.rar' '*.zip' '*.info' '*.torrent' '*.sample*' '*.proof*'
@@ -660,7 +799,6 @@ MEDIA_MAINTENANCE_JOBS=(
LIDARR_MUSIC_ROOT="/mnt/user/Music-New" # must match the root path set in Lidarr
LIDARR_ORPHAN_AGE=7 # days before untracked file is eligible for deletion
LIDARR_EXTENSIONS=("flac" "mp3" "m4a" "wav" "aac" "ogg" "opus" "wma")
# file extensions considered valid music files
LIDARR_PROTECTED_PATTERNS=("*.jpg" "*.jpeg" "*.png" "*.nfo" "*.lrc")
# never deleted — cover art, metadata, lyrics
@@ -684,37 +822,26 @@ MEDIA_MAINTENANCE_JOBS=(
# ==============================================================================================
# ── TRANSCODES ────────────────────────────────────────────────────────────────────────────────
# ==============================================================================================
# Session-based storage allocator using filesystem symlink indirection.
# ffmpeg resolves the symlink ONCE at session start — existing sessions are never affected.
# Only new sessions care about where the symlink currently points.
#
# How it works:
# ramdisk_setup.sh — run once at array start, creates tmpfs and sets symlink
# transcode_manager.sh — every 3 min, monitors usage and manages symlink direction
# transcode_cleanup.sh — every 5 min, removes old inactive files from both locations
#
# ⚠️ Docker mount warning:
# Do NOT add a static SSD transcode path as a second volume mount in your Emby container.
# If the SSD path is mounted inside the container Emby can see it and will use it
# independently of the symlink — breaking symlink-based routing entirely.
# The symlink IS your emergency lever — one mount only:
# /mnt/ram-transcode → /ext-ram-transcode
# ⚠️ One mount only in Emby container: /mnt/ram-transcode → /ext-ram-transcode
# Do NOT add a static SSD path — Emby will use it independently of the symlink.
# The symlink IS your emergency lever — flip it manually if needed:
# ln -sfn /mnt/cache/Temp_Storage/Emby/Transcodes /mnt/ram-transcode
RAMDISK_PATH="/mnt/ramdisk_transcodes" # tmpfs mount point created at array start
RAMDISK_SIZE="8G" # ceiling — tmpfs only uses RAM actually needed
TRANSCODE_LINK="/mnt/ram-transcode" # symlink Emby points at — location never changes
TRANSCODE_SSD="/mnt/cache/Temp_Storage/Emby/Transcodes/" # SSD fallback location
# Usage thresholds in GB — hysteresis gap prevents flip-flop near threshold
# Usage thresholds in GB — hysteresis gap between WARN and LOW prevents flip-flop
RAMDISK_WARN_GB=6.8 # flip symlink to SSD at or above this usage
RAMDISK_LOW_GB=5.5 # flip symlink back to ramdisk when usage drops here
RAMDISK_SSD_MIN_GB=20 # minimum free GB on SSD required before allowing flip to SSD
# Cleanup age thresholds — files must be older than these AND not open by any process
TRANSCODE_MAX_AGE=20 # minutes before a transcode file is eligible for cleanup
TRANSCODE_ORPHAN_AGE=30 # minutes before an orphaned file is eligible
TRANSCODE_ORPHAN_AGE=30 # minutes before an orphaned file is eligible — extra caution buffer
# Flip frequency alert — too many flips per hour may indicate ramdisk needs to be larger
# Flip frequency alert — too many flips per hour indicates ramdisk needs to be larger
TRANSCODE_FLIP_WARN=3 # notify if symlink flips this many times in one hour
# Permissions — must match your Emby container user
@@ -729,28 +856,23 @@ MEDIA_MAINTENANCE_JOBS=(
# ssd — always uses SSD, never uses ramdisk
# useful during ramdisk maintenance, testing, or after a flip issue
# switch to this mode to drain ramdisk sessions gracefully
TRANSCODE_MANAGER_MODE="smart"
TRANSCODE_MANAGER_MODE="smart" # smart | ramdisk | ssd
# Emby container check — skips threshold checks when Emby is not running
# Prevents unnecessary symlink flips when no transcoding is happening
TRANSCODE_CHECK_EMBY=true
TRANSCODE_EMBY_CONTAINER="Emby" # exact Docker container name — case sensitive
TRANSCODE_EMBY_CONTAINER="Emby" # exact Docker container name — case sensitive
# ==============================================================================================
# ── MONITORS ──────────────────────────────────────────────────────────────────────────────────
# ── MONITORS ──────────────────────────────────────────────────────────────────────────────────
# ==============================================================================================
# Monitoring scripts — watch and report only, never take action.
# Lives in Monitor/ folder — distinct from unRAID_Essentials (which acts) and
# Docker_Essentials (which manages containers).
# These scripts are the data sources for the future plugin dashboard.
# ━━━ Certificate Monitor ━━━
# Checks SSL cert expiry via direct openssl connection — no NPM dependency.
# Reads the actual cert the server is presenting — catches real-world issues API checks miss.
# Each domain and subdomain is a separate entry — they have independent certs.
# Add your public-facing domains — uncomment and replace with your actual domains.
CERT_MONITOR_DOMAINS=(
"Gmer4Lfe.com"
"Gmer4Lfe.com"
"Gmer4Lfe.us"
)
CERT_WARN_DAYS=30 # notify warning when cert expires within this many days
@@ -763,9 +885,6 @@ CERT_MONITOR_DOMAINS=(
# Leave BACKUP_VERIFY_SHARES empty to automatically use DAILY_SYNC_SHARES as the target list.
BACKUP_VERIFY_SHARES=(
# leave empty to use DAILY_SYNC_SHARES automatically
# or specify individual shares to verify:
# /mnt/user/Movies
# /mnt/user/Tv_Shows
)
BACKUP_VERIFY_SAMPLE=10 # number of files to randomly sample per share per run
BACKUP_VERIFY_MIN_SIZE=1M # skip files smaller than this — avoids tiny junk files
@@ -777,49 +896,45 @@ BACKUP_VERIFY_SHARES=(
SMART_TEMP_WARN=45 # degrees C — warn if drive temperature exceeds this
SMART_TEMP_CRIT=55 # degrees C — critical if drive temperature exceeds this
SMART_IGNORE_DRIVES=(
"sda" # uncomment to ignore sda — common choice if sda is your unRAID boot USB
"sda" # boot USB — SMART not meaningful on flash drives
)
# ━━━ Bandwidth Monitor ━━━
# Logs daily rsync transfer totals to a bounded file on /boot/ — minimal flash wear.
# bandwidth_monitor.sh --log-transfer is called by rsync.sh after each successful sync.
# bandwidth_monitor.sh --report generates the weekly summary standalone.
# File stays bounded to BANDWIDTH_LOG_RETENTION lines — old entries auto-purged on each write.
BANDWIDTH_LOG="/boot/config/bandwidth_history.db"
BANDWIDTH_LOG_RETENTION=90 # days to keep — file never grows beyond ~90 lines
BANDWIDTH_WARN_GB=50 # flag in reports if a single sync transfer exceeds this GB
# ━━━ ZFS Memory Snapshot ━━━
# ━━━ ZFS Memory Snapshot ━━━
# Weekly ZFS pool health and memory diagnostic report — informational only, no action taken.
# system_watchdog.sh handles threshold-based intervention.
# Output written to ZFS_REPORT_LOG for historical review in addition to console output.
# Pools in ZFS_REPORT_IGNORE_POOLS are excluded from reporting — still monitored by unRAID.
ZFS_REPORT_LOG="/var/log/zfs-weekly-health.log"
ZFS_REPORT_ARC_WARN_PCT=90
ZFS_REPORT_FREE_WARN_GB=10
ZFS_REPORT_AVAIL_WARN_GB=20
ZFS_REPORT_DOCKER_TOP=10
# Pools to exclude from health reporting — still monitored by unRAID but skipped in report
# Useful for pools that are expected to be heavily used or are managed separately
ZFS_REPORT_ARC_WARN_PCT=90 # warn in report if ARC utilization above this %
ZFS_REPORT_FREE_WARN_GB=10 # warn in report if free RAM drops below this GB
ZFS_REPORT_AVAIL_WARN_GB=20 # warn in report if available RAM drops below this GB
ZFS_REPORT_DOCKER_TOP=10 # number of top Docker memory users to show in report
ZFS_REPORT_IGNORE_POOLS=(
"disk10" # Docker overlay storage — high usage is normal
"disk9" # Cache pool — usage varies widely, not meaningful to report
# Pools excluded from health reporting — expected to run at high usage
# All pools still monitored by unRAID regardless of this list
"disk10"
"disk9"
"disk8"
"disk6"
"disk5"
)
# ━━━ Bandwidth Monitor ━━━
# Called by rsync.sh after each sync — one bounded write per run, minimal flash wear.
# Log format: YYYY-MM-DD|HH:MM|profile|duration_seconds|status — version-proof
# File stays bounded to BANDWIDTH_LOG_RETENTION days — old entries auto-purged on write.
BANDWIDTH_LOG="/boot/config/bandwidth_history.db"
BANDWIDTH_LOG_RETENTION=90 # days to keep — file never grows beyond ~90 lines
BANDWIDTH_WARN_GB=50 # flag in reports if a single sync transfer exceeds this GB
# ━━━ Health Digest ━━━
# Aggregated system health summary from across the ecosystem.
# Reads existing state files — no new writes to flash drive.
#
# Aggregated system health summary — reads existing state files, no new writes to flash.
# Three profiles — switch by changing DIGEST_PROFILE, no cron changes needed:
# always — sends every run (schedule daily = daily digest, weekly = weekly digest)
# smart — sends only if findings worth reporting (intelligent quiet operation)
# weekly — sends once per week on DIGEST_DAY only, silent all other days
#
# Data sources (reads only — no writes):
# Transcode ramdisk state, container watchdog strikes, system watchdog strikes,
# failover state, container skip list, bandwidth history, SSL cert days remaining
# always — sends every run
# smart — sends only if findings worth reporting
# weekly — sends once per week on DIGEST_DAY only
DIGEST_PROFILE="weekly" # always | smart | weekly
DIGEST_DAY="Sunday" # day name for weekly profile — must match date +%A output
DIGEST_DAY="Sunday" # must match date +%A output
# Smart profile triggers — set true to send digest when this condition is found
DIGEST_SMART_ON_WATCHDOG=true # send if any watchdog strikes are active
@@ -828,29 +943,19 @@ ZFS_REPORT_IGNORE_POOLS=(
DIGEST_SMART_ON_BANDWIDTH=true # send if any transfer exceeded BANDWIDTH_WARN_GB
# ━━━ Emby Session Report ━━━
# Weekly Emby usage report via API — no persistent writes, queries fresh each run.
# Shows active streams, library counts, transcode vs direct play ratio.
# Requires an API key from Emby Settings → API Keys in the Emby WebUI.
# Requires API key from Emby Settings → API Keys in the Emby WebUI.
# No persistent writes — queries fresh each run.
EMBY_URL="http://localhost:8096"
EMBY_API_KEY="0c27448d93a7431f9ac63569f7655829" # paste your Emby API key here
EMBY_API_KEY="0c27448d93a7431f9ac63569f7655829"
EMBY_REPORT_DAYS=7 # number of days to include in the report period
EMBY_REPORT_TOP_N=10 # number of top content items to show in report
# ==============================================================================================
# ── SYSTEM WATCHDOG ───────────────────────────────────────────────────────────────────────────
# ==============================================================================================
# Last line of defense — reboots the system cleanly when it is about to become unstable.
# Runs every 15 minutes via cron. Works alongside docker_watchdog.sh:
# docker_watchdog.sh — container level, minimal disruption, tries to self-heal first
# system_watchdog.sh — system level, last resort, reboots when healing has failed
#
# Strike system: sustained threshold hits trigger reboot — single spikes are ignored.
# Each check that exceeds its threshold adds a strike. Strikes reset when recovered.
# When strike limit is hit the reboot sequence begins.
#
# Reboot loop protection: tracks reboot timestamps on /boot/ (survives reboots).
# If the server reboots too many times in the window it shuts down instead — a reboot
# loop means something is fundamentally wrong that a reboot is not fixing.
# Last line of defense — reboots cleanly when system is about to become unstable.
# Strike system: sustained threshold hits trigger reboot — single spikes ignored.
# Reboot loop protection: shuts down instead if reboot limit hit in window.
# ━━━ State Files ━━━
# Strike counts reset on reboot — /tmp is correct (fresh start after each reboot)