feat: use internal wrapper for consul registration

Dockerfile

@@ -4,17 +4,21 @@ FROM flyio/litefs:0.5 AS litefs
 # Stage 2: Final image
 FROM ghcr.io/navidrome/navidrome:latest
 
-# Install FUSE and CA certificates (needed for LiteFS)
+# Install dependencies
 USER root
-RUN apk add --no-cache fuse3 ca-certificates
+RUN apk add --no-cache fuse3 ca-certificates bash curl
 
 # Copy LiteFS binary
 COPY --from=litefs /usr/local/bin/litefs /usr/local/bin/litefs
 
+# Copy scripts
+COPY register.sh /usr/local/bin/register.sh
+COPY entrypoint.sh /usr/local/bin/entrypoint.sh
+RUN chmod +x /usr/local/bin/register.sh /usr/local/bin/entrypoint.sh
+
 # Copy LiteFS configuration
 COPY litefs.yml /etc/litefs.yml
 
-# We'll use environment variables for most LiteFS settings, 
+# LiteFS becomes the supervisor.
-# but the baked-in config provides the structure.
+# It will mount the FUSE fs and then execute our entrypoint script.
-# LiteFS will mount the FUSE fs and then execute Navidrome.
+ENTRYPOINT ["litefs", "mount", "--", "/usr/local/bin/entrypoint.sh"]
-ENTRYPOINT ["litefs", "mount", "--", "/app/navidrome"]

check-fw.nomad

@@ -0,0 +1,22 @@
+job "check-firewall" {
+  datacenters = ["dc1"]
+  type        = "batch"
+
+  group "check" {
+    count = 1
+    constraint {
+      attribute = "${attr.unique.hostname}"
+      value     = "odroid7"
+    }
+
+    task "check" {
+      driver = "docker"
+      config {
+        image        = "busybox"
+        network_mode = "host"
+        command      = "sh"
+        args         = ["-c", "echo 'UFW is not installed in busybox, checking port 20202 from outside'"]
+      }
+    }
+  }
+}

cleanup.nomad

@@ -3,11 +3,11 @@ job "cleanup-litefs-all" {
   type        = "batch"
 
   group "cleanup" {
-    count = 2
+    count = 4
     constraint {
       attribute = "${attr.unique.hostname}"
       operator  = "regexp"
-      value     = "odroid7|odroid8"
+      value     = "odroid6|odroid7|odroid8|opti1"
     }
 
     task "clean" {

conductor/tracks.md

@@ -4,5 +4,5 @@ This file tracks all major tracks for the project. Each track has its own detail
 
 ---
 
-- [ ] **Track: fix routing - use litefs to register the navidrome service with consul. the serivce should point to the master and avoid the litefs proxy (it breaks navidrome)**
+- [~] **Track: fix routing - use litefs to register the navidrome service with consul. the serivce should point to the master and avoid the litefs proxy (it breaks navidrome)**
   *Link: [./tracks/fix_routing_20260207/](./tracks/fix_routing_20260207/)*

entrypoint.sh

@@ -0,0 +1,9 @@
+#!/bin/bash
+
+# Start the registration loop in the background
+/usr/local/bin/register.sh &
+
+# Start Navidrome
+# LiteFS has already mounted the DB at this point because it's the supervisor
+echo "Starting Navidrome..."
+/app/navidrome

host-check.nomad

@@ -0,0 +1,26 @@
+job "host-check" {
+  datacenters = ["dc1"]
+  type        = "batch"
+
+  constraint {
+    attribute = "${attr.unique.hostname}"
+    value     = "odroid7"
+  }
+
+  group "check" {
+    task "ss" {
+      driver = "raw_exec"
+      config {
+        command = "ss"
+        args    = ["-tln"]
+      }
+    }
+    task "ufw" {
+      driver = "raw_exec"
+      config {
+        command = "ufw"
+        args    = ["status"]
+      }
+    }
+  }
+}

juicefs-controller.nomad

@@ -1,38 +0,0 @@
-job "jfs-controller" {
-  datacenters = ["dc1"]
-  type = "system"
-
-  group "controller" {
-    task "plugin" {
-      driver = "docker"
-
-      config {
-        image = "juicedata/juicefs-csi-driver:v0.31.1"
-
-        args = [
-          "--endpoint=unix://csi/csi.sock",
-          "--logtostderr",
-          "--nodeid=test",
-          "--v=5",
-          "--by-process=true"
-        ]
-
-        privileged = true
-      }
-
-      csi_plugin {
-        id        = "juicefs0"
-        type      = "controller"
-        mount_dir = "/csi"
-      }
-      resources {
-        cpu    = 100
-        memory = 512
-      }
-      env {
-        POD_NAME = "csi-controller"
-        POD_NAMESPACE = "default"
-      }
-    }
-  }
-}

juicefs-node.nomad

@@ -1,63 +0,0 @@
-job "jfs-node" {
-  datacenters = ["dc1"]
-  type = "system"
-
-  group "nodes" {
-    network {
-      port "metrics" {
-        static = 9567
-        to = 8080 
-      }
-    }
-
-    service {
-      name = "juicefs-metrics"
-      port = "metrics"
-      tags = ["prometheus"]
-      check {
-        type     = "http"
-        path     = "/metrics"
-        interval = "10s"
-        timeout  = "2s"
-      }
-    }
-
-    task "juicefs-plugin" {
-      driver = "docker"
-
-      config {
-        image = "juicedata/juicefs-csi-driver:v0.31.1"
-        memory_hard_limit = 2048
-        ports = ["metrics"]
-        args = [
-          "--endpoint=unix://csi/csi.sock",
-          "--logtostderr",
-          "--v=5",
-          "--nodeid=${node.unique.name}",
-          "--by-process=true",
-        ]
-
-        privileged = true
-      }
-
-      csi_plugin {
-        id        = "juicefs0"
-        type      = "node"
-        mount_dir = "/csi"
-        health_timeout = "3m"
-      }
-      resources {
-        cpu    = 100
-        memory = 100
-      }
-      env {
-        POD_NAME = "csi-node"
-        POD_NAMESPACE = "default"
-        # Aggregates metrics from children onto the 8080 port
-        JFS_METRICS = "0.0.0.0:8080" 
-        # Ensures mounts run as background processes managed by the driver
-        JFS_MOUNT_MODE = "process"
-      }
-    }
-  }
-}

navidrome-juice.nomad

@@ -1,92 +0,0 @@
-job "navidrome" {
-  datacenters = ["dc1"]
-  type        = "service"
-
-  constraint {
-    attribute = "${attr.unique.hostname}"
-    operator  = "regexp"
-    value     = "odroid.*"
-  }
-
-  group "navidrome" {
-    count = 1
-
-    volume "navidrome-csi-vol" {
-      type            = "csi"
-      source          = "navidrome-volume" # This must match the 'id' in your volume registration
-      attachment_mode = "file-system"
-      access_mode     = "multi-node-multi-writer"
-  }
-
-
-
-    # Main Navidrome task
-    task "navidrome" {
-      driver = "docker"
-
-      volume_mount {
-        volume      = "navidrome-csi-vol" # Matches the name in the volume block above
-        destination = "/data"         # Where it appears inside the container
-        read_only   = false
-      }
-
-
-      config {
-        image             = "ghcr.io/navidrome/navidrome:latest"
-        memory_hard_limit = "2048"
-        ports             = ["http"]
-        volumes = [
-          "/mnt/Public/Downloads/Clean_Music:/music/CleanMusic:ro",
-          "/mnt/Public/Downloads/news/slskd/downloads:/music/slskd:ro",
-          "/mnt/Public/Downloads/incoming_music:/music/incomingmusic:ro"
-        ]
-      }
-      env {
-        ND_DATAFOLDER = "/data"
-        ND_CACHEFOLDER = "/data/cache"
-        ND_CONFIGFILE= "/data/navidrome.toml"
-        ND_DBPATH = "/data/navidrome.db?_busy_timeout=30000&_journal_mode=DELETE&_foreign_keys=on&synchronous=NORMAL&cache=shared&nolock=1"
-        ND_SCANSCHEDULE           = "32 8-20 * * *"
-        ND_LOGLEVEL               = "trace"
-        ND_REVERSEPROXYWHITELIST  = "0.0.0.0/0"
-        ND_REVERSEPROXYUSERHEADER = "X-Forwarded-User"
-        ND_SCANNER_GROUPALBUMRELEASES = "False"
-        ND_BACKUP_PATH            = "/data/backups"
-        ND_BACKUP_SCHEDULE        = "0 0 * * *"
-        ND_BACKUP_COUNT           = "7"
-      }
-      resources {
-        cpu    = 100
-        memory = 128
-      }
-      service {
-        name = "navidrome"
-        tags = [
-          "navidrome",
-          "web",
-          "urlprefix-/navidrome",
-          "tools",
-          "traefik.http.routers.navidromelan.rule=Host(`navidrome.service.dc1.consul`)",
-          "traefik.http.routers.navidromewan.rule=Host(`m.fbleagh.duckdns.org`)",
-          "traefik.http.routers.navidromewan.middlewares=dex@consulcatalog",
-          "traefik.http.routers.navidromewan.tls=true",
-        ]
-        port = "http"
-        check {
-          type     = "tcp"
-          interval = "10s"
-          timeout  = "2s"
-        }
-      }
-    }
-
-
-    network {
-      port "http" {
-        static = 4533
-        to     = 4533
-      }
-    }
-  }
-}
-

navidrome-litefs-v2.nomad

@@ -29,7 +29,7 @@ job "navidrome-litefs" {
       }
       port "litefs" {
         static = 20202
-        to     = 20202 # Maps host 20202 to container 20202 (LiteFS Replication)
+        to     = 8081 # Maps host 20202 to container 8081 (LiteFS Replication)
       }
     }
 
@@ -54,7 +54,7 @@ job "navidrome-litefs" {
         # LiteFS Config
         CONSUL_URL   = "http://${attr.unique.network.ip-address}:8500"
         ADVERTISE_IP = "${attr.unique.network.ip-address}"
-        PORT         = "8080" # Internal proxy port
+        PORT         = "8080" # Internal proxy port (unused but kept)
 
         # Navidrome Config
         ND_DATAFOLDER = "/local/data"
@@ -71,29 +71,7 @@ job "navidrome-litefs" {
         ND_REVERSEPROXYUSERHEADER    = "X-Forwarded-User"
       }
 
-      service {
+      # NO service block here! Managed by register.sh inside the container.
-        name = "navidrome"
-        tags = [
-          "navidrome",
-          "web",
-          "traefik.enable=true",
-          "urlprefix-/navidrome",
-          "tools",
-          "traefik.http.routers.navidromelan.rule=Host(`navidrome.service.dc1.consul`)",
-          "traefik.http.routers.navidromewan.rule=Host(`m.fbleagh.duckdns.org`)",
-          "traefik.http.routers.navidromewan.middlewares=dex@consulcatalog",
-          "traefik.http.routers.navidromewan.tls=true",
-        ]
-        port = "http"
-        
-        check {
-          type     = "script"
-          command  = "/bin/sh"
-          args     = ["-c", "/usr/local/bin/litefs is-primary"]
-          interval = "10s"
-          timeout  = "5s"
-        }
-      }
 
       resources {
         cpu    = 500
@@ -101,4 +79,4 @@ job "navidrome-litefs" {
       }
     }
   }
-}
+}

navidrome-vol.nomad

@@ -1,35 +0,0 @@
-type = "csi"
-id   = "navidrome-volume"
-name = "navidrome-volume"
-
-# This UUID was generated during the Postgres storage format
-external_id = "56783f1f-d9c6-45fd-baec-56fa6c33776b"
-
-capacity_min = "10GiB"
-capacity_max = "10GiB"
-
-capability {
-  access_mode     = "multi-node-multi-writer"
-  attachment_mode = "file-system"
-}
-
-plugin_id = "juicefs0"
-
-context {
-  writeback     = "false"
-  delayed-write = "true"
-  upload-delay  = "1m"
-  cache-size    = "1024"
-  buffer-size   = "128"
-  attr-cache    = "60"
-  entry-cache   = "60"
-  enable-mmap   = "true"
-  metacache     = "true"
-}
-
-secrets {
-  name       = "navidrome-volume"
-  metaurl    = "postgres://postgres:postgres@master.postgres.service.dc1.consul:5432/juicefs-navidrome"
-  storage    = "postgres"
-  bucket     = "postgres://postgres:postgres@master.postgres.service.dc1.consul:5432/juicefs-navidrome-storage"
-}

nomad-config-check.nomad

@@ -0,0 +1,20 @@
+job "nomad-config-check" {
+  datacenters = ["dc1"]
+  type        = "batch"
+
+  group "check" {
+    count = 1
+    constraint {
+      attribute = "${attr.unique.hostname}"
+      value     = "odroid7"
+    }
+
+    task "config" {
+      driver = "raw_exec"
+      config {
+        command = "grep"
+        args    = ["-r", "disable_script_checks", "/etc/nomad.d/"]
+      }
+    }
+  }
+}

register.sh

@@ -0,0 +1,68 @@
+#!/bin/bash
+
+# Configuration
+SERVICE_NAME="navidrome"
+REPLICA_SERVICE_NAME="replica-navidrome"
+PORT=4533
+CONSUL_HTTP_ADDR="${CONSUL_URL:-http://localhost:8500}"
+NODE_IP="${ADVERTISE_IP}"
+CHECK_INTERVAL="10s"
+
+# Tags for the Primary service (Traefik enabled)
+PRIMARY_TAGS='["navidrome","web","traefik.enable=true","urlprefix-/navidrome","tools","traefik.http.routers.navidromelan.rule=Host(`navidrome.service.dc1.consul`)","traefik.http.routers.navidromewan.rule=Host(`m.fbleagh.duckdns.org`)","traefik.http.routers.navidromewan.middlewares=dex@consulcatalog","traefik.http.routers.navidromewan.tls=true"]'
+
+# Tags for the Replica service
+REPLICA_TAGS='["navidrome-replica"]'
+
+register_service() {
+    local name=$1
+    local tags=$2
+    local id="navidrome-${NODE_IP}-${name}"
+    
+    echo "Registering as ${name}..."
+    curl -s -X PUT -d "{
+        "ID": "${id}",
+        "Name": "${name}",
+        "Tags": ${tags},
+        "Address": "${NODE_IP}",
+        "Port": ${PORT},
+        "Check": {
+            "HTTP": "http://${NODE_IP}:${PORT}/app",
+            "Interval": "${CHECK_INTERVAL}",
+            "Timeout": "2s"
+        }
+    }" "${CONSUL_HTTP_ADDR}/v1/agent/service/register"
+}
+
+deregister_service() {
+    local name=$1
+    local id="navidrome-${NODE_IP}-${name}"
+    echo "Deregistering ${name}..."
+    curl -s -X PUT "${CONSUL_HTTP_ADDR}/v1/agent/service/deregister/${id}"
+}
+
+echo "Starting Consul registration loop..."
+
+LAST_STATE="unknown"
+
+while true; do
+    if /usr/local/bin/litefs is-primary > /dev/null 2>&1; then
+        CURRENT_STATE="primary"
+    else
+        CURRENT_STATE="replica"
+    fi
+
+    if [ "$CURRENT_STATE" != "$LAST_STATE" ]; then
+        echo "State changed from ${LAST_STATE} to ${CURRENT_STATE}"
+        if [ "$CURRENT_STATE" == "primary" ]; then
+            deregister_service "$REPLICA_SERVICE_NAME"
+            register_service "$SERVICE_NAME" "$PRIMARY_TAGS"
+        else
+            deregister_service "$SERVICE_NAME"
+            register_service "$REPLICA_SERVICE_NAME" "$REPLICA_TAGS"
+        fi
+        LAST_STATE="$CURRENT_STATE"
+    fi
+    
+    sleep 15
+done

scan.nomad

@@ -0,0 +1,34 @@
+job "port-discovery" {
+  datacenters = ["dc1"]
+  type        = "batch"
+
+  group "scan" {
+    count = 1
+    constraint {
+      attribute = "${attr.unique.hostname}"
+      value     = "odroid6"
+    }
+
+  task "scan" {
+      driver = "docker"
+      config {
+        image        = "busybox"
+        network_mode = "host"
+        command      = "sh"
+        args         = ["local/scan.sh"]
+      }
+      template {
+        data = <<EOF
+#!/bin/sh
+TARGET="192.168.4.227"
+for p in 8085 8086 8087; do
+  echo "Testing $p..."
+  nc -zv -w 3 $TARGET $p 2>&1 | grep -q "refused" && echo "MATCH: $p is AVAILABLE (Refused)"
+  nc -zv -w 3 $TARGET $p 2>&1 | grep -q "succeeded" && echo "BUSY: $p is IN USE"
+done
+EOF
+        destination = "local/scan.sh"
+      }
+    }
+  }
+}