fix: prefer real block devices over tmpfs for storage display

The disk-collector could produce an empty fsSize array when
/host/proc/1/mounts is unreadable, causing the admin UI to fall back
to systeminformation's fsSize which includes tmpfs mounts. This led to
the storage display showing ~1.5 GB (tmpfs /run) instead of the actual
storage capacity.

Two changes:
- disk-collector: fall back to df on /storage when host mount table
  yields no real filesystems, since /storage is always bind-mounted
  from the host and reflects the actual backing device.
- easy-setup UI: when falling back to systeminformation fsSize, filter
  for /dev/ block devices and prefer the largest one instead of blindly
  taking the first entry.

Fixes #373

admin/inertia/pages/easy-setup/index.tsx

@@ -318,7 +318,19 @@ export default function EasySetupWizard(props: { system: { services: ServiceSlim
   }
 
   const primaryDisk = getPrimaryDisk()
-  const primaryFs = systemInfo?.fsSize?.[0]
+  // When falling back to fsSize (systeminformation), prefer real block devices
+  // over virtual filesystems like tmpfs which report misleading capacity.
+  const getPrimaryFs = () => {
+    if (!systemInfo?.fsSize || systemInfo.fsSize.length === 0) return null
+    const realDevices = systemInfo.fsSize.filter((fs) => fs.fs.startsWith('/dev/'))
+    if (realDevices.length > 0) {
+      return realDevices.reduce((largest, current) =>
+        current.size > largest.size ? current : largest
+      )
+    }
+    return systemInfo.fsSize[0]
+  }
+  const primaryFs = getPrimaryFs()
   const storageInfo = primaryDisk
     ? { totalSize: primaryDisk.totalSize, totalUsed: primaryDisk.totalUsed }
     : primaryFs

install/sidecar-disk-collector/collect-disk-info.sh

@@ -50,6 +50,22 @@ while true; do
         FS_JSON+="{\"fs\":\"${dev}\",\"size\":${size},\"used\":${used},\"available\":${avail},\"use\":${pct},\"mount\":\"${mountpoint}\"}"
         FIRST=0
     done < /host/proc/1/mounts
+
+    # Fallback: if no real filesystems were found from the host mount table
+    # (e.g. /host/proc/1/mounts was unreadable), try the /storage mount directly.
+    # The disk-collector container always has /storage bind-mounted from the host,
+    # so df on /storage reflects the actual backing device and its capacity.
+    if [[ "$FIRST" -eq 1 ]] && mountpoint -q /storage 2>/dev/null; then
+        STATS=$(df -B1 /storage 2>/dev/null | awk 'NR==2{print $1,$2,$3,$4,$5}')
+        if [[ -n "$STATS" ]]; then
+            read -r dev size used avail pct <<< "$STATS"
+            pct="${pct/\%/}"
+            FS_JSON+="{\"fs\":\"${dev}\",\"size\":${size},\"used\":${used},\"available\":${avail},\"use\":${pct},\"mount\":\"/storage\"}"
+            FIRST=0
+            log "Used /storage mount as fallback for filesystem info."
+        fi
+    fi
+
     FS_JSON+="]"
 
     # Use a tmp file for atomic update