Cloudplow - Cloudbox/Cloudbox GitHub Wiki

Intro

Cloudplow (CP) is a script created by l3uddz that has two main components (as related to Cloudbox):

  1. Uploader to Rclone remote: Files are moved off local storage. With support for multiple uploaders (i.e. remote/folder pairings).

  2. UnionFS Cleaner functionality: Deletion of UnionFS-Fuse whiteout files (*_HIDDEN~) and their corresponding "whited-out" files on Rclone remotes. With support for multiple remotes (useful if you have multiple Rclone remotes mounted). Cloudbox has switched to using mergerfs by default for the merged file system at /mnt/unionfs, so this section no longer applies to the default setup. It's left here for reference.

Remote Uploader Function

As setup for Cloudbox, Cloudplow uploads all the content in /mnt/local/Media/ (see Paths) to your cloud storage provider (e.g. Google Drive), after the folder reaches a 200 GB size threshold, when checked every 30 minutes.

Note: The size threshold and the check interval can be changed via steps mentioned on this page.

Google Drive Daily Upload Limit (click to expand)

Recently, Google Drive has a max upload limit of about 750GB per day. When this limit is reached, Google Drive will put you in a 24 hour soft ban. When Cloudplow detects this (with the phrase Failed to copy: googleapi: Error 403: User rate limit exceeded), uploading will be suspended for 25 hours (i.e. a 25 hour ban sleep), and upon waking up, it will resume its checking and uploading tasks. This feature is enabled by default. This method is better than running Rclone task with a bwlimit, becasuse you can just upload in bursts when the uploading resumes.

Note: The keywords or phrases that are used to monitor the ban, and the duration of the sleep time, can be changed at any time by editing the config.json file.

UnionFS Cleaner Function

Cloudbox has switched to using mergerfs by default for the merged file system at /mnt/unionfs, so this section no longer applies to the default setup. It's left here for reference.

On top of uploading your media files to the cloud storage, Cloudplow also functions as a UnionFS whiteout cleaner (i.e. *_HIDDEN~ files).

What are *_HIDDEN~ files? (click to expand)

When Sonarr & Radarr upgrade your media files, they attempt delete the previous ones. When that data is still on the local server, it is deleted immediately, but when it has been moved to the cloud storage provider, for example Google Drive, it is unable to do so because the Google Drive mount is set as read-only (via Plexdrive or Rclone VFS).

Instead, UnionFS will create a whiteout file (a blank file in the format of filename.ext_HIDDEN~), at /mnt/local/.unionfs-fuse/ and that will make the file invisible to whatever tries to access it via the UnionFS mount (.e.g. /mnt/unionfs/) and, therefore, Sonarr & Radarr will consider the file deleted, however, the media file will still exist on the cloud.

To resolve this, on the next upload task (i.e. when size threshold is reached on the next interval check), Cloudplow will scan for the whiteout file(s), remove the corresponding media file from the cloud storage, then remove the whiteout file (since it isn't needed anymore), and as a result, keep your content free of duplicates.

Config

Default config.json file

See Config: Cloudplow.

Location

/opt/cloudplow/config.json

Note: Config changes require a restart: sudo systemctl restart cloudplow.

Editing

Edit in your favorite code editor (with json highlighting) or even a unix editor like nano.

nano /opt/cloudplow/config.json

Note: The cloudplow config file is a JSON file. JSON files have a particular format and syntax. If you are unfamiliar with JSON formatting and syntax, don't edit this file until you have gained that familiarity. Here's a random YouTube video that will give you a ten-minute overview.

Modify Upload Threshold and Interval

    "uploader": {
        "google": {
            "check_interval": 30,
            "exclude_open_files": false,
            "max_size_gb": 200,
            "opened_excludes": [
                "/downloads/"
            ],
            "size_excludes": [
                "downloads/*"
            ]
        }

"check_interval": How often (in minutes) Cloudplow checks the size of /mnt/local/Media.

"max_size_gb": Max size (in GB) Cloudplow allows /mnt/local/Media to get before starting an upload task.

  • Note: max_size_gb is rounded up, so it is advised to have it minimum 2GB or else it would attempt upload at each interval. Explanation below.

    • 1GB is basically anything in there.

    • 2GB is at least 1GB of data.

Plex Integration

Cloudplow can throttle Rclone uploads during active, playing Plex streams (paused streams are ignored).

  "plex": {
      "enabled": false,
      "url": "https://plex.domain.com",
      "token": "",
      "poll_interval": 60,
      "max_streams_before_throttle": 1,
      "rclone": {
          "throttle_speeds": {
              "0": "1000M",
              "1": "50M",
              "2": "40M",
              "3": "30M",
              "4": "20M",
              "5": "10M"
          },
          "url": "http://localhost:7949"
      }
  }

enabled - Change false to true to enable.

url - Your Plex URL.

token - Your Plex Access Token.

poll_interval - How often (in seconds) Plex is checked for active streams.

max_streams_before_throttle - How many playing streams are allowed before enabling throttling.

rclone

  • url - Leave as default.

  • throttle_speeds - Categorized option to configure upload speeds for various stream counts (where 5 represents 5 or more streams). M is MB/s.

    • Format:

      "STREAM COUNT": "THROTTLED UPLOAD SPEED",

Pushover Notifications

See here.

Restart

Restart Cloudplow to apply the changes to the config.

sudo systemctl restart cloudplow

CLI

You can run a manual Cloudplow task from anywhere by just using the cloudplow command.

Manual Clean

To clean the hidden files and remove deleted files from the cloud:

cloudplow clean

Manual Upload

To start uploading right away, regardless of what the folder size is:

cloudplow upload

750GB Google Drive limit

Google Drive has a 750GB/day upload limit. Cloudplow will automatically detect this and put downloading into a 25-hour timeout. There are ways around this using service accounts, but they are beyond the scope of this guide. You can find out details via the Discord.

The triggers for this timeout are in the rclone_sleeps section of the config, and more details can be found here.

⚠️ **GitHub.com Fallback** ⚠️