Rclone as a helper for external backups

Rsync for the Cloud

Article from Issue 227/2019
Author(s):

Backups for critical data have to withstand major disasters. Rclone, a descendant of Rsync, enables easy backups to the cloud and supports many services.

We all live in the hope of never getting into a situation where our belongings are destroyed by theft, fire, water, or natural disaster, but there are no guarantees. If you are serious about preserving your important digital data, you should secure it outside your home, preferably in several places at the same time. The Rsync-based Rclone [1] tool helps with off-site backup scenarios.

Rclone is a command-line program that performs one-way synchronization between locally defined datasets and a cloud. If you run it a second time, it searches the local filesystem for changes and then incrementally uploads only the changes to the cloud of your choice. Rclone implements most of Rsync's options and syntax. In addition, it offers additional commands to optimally support the individual services. Rclone offers a graphical front end for people who prefer not to work at the command line.

Rsync's two-way synchronization is not part of Rclone's feature set. This means that it deletes any locally-deleted files from the cloud, but not vice versa. In addition, the program does not save older versions when changes are made to files. It overwrites changed files, unless the selected service has its own versioning system, such as Google Drive or Dropbox. One way to avoid this is to use the backup-dir function, which moves changed or deleted files to a separate directory [2].

The tool, written in Go, which can be found in the archives of most distributions, supports numerous cloud service providers and protocols, including the open source clouds Nextcloud, ownCloud, and OpenStack Swift, as well as the commercial offerings Amazon S3, Box, Dropbox, Google Drive, pCloud, and Microsoft OneDrive [3].

Rclone also supports the HTTP, FTP, SFTP, and WebDAV protocols, so web and file servers, as well as NAS devices, can be integrated as back ends using standard protocols. The program is available in variants for x86, ARM, and MIPS architectures. Not only for Linux, Rclone also is available for macOS, Windows, BSD, Plan 9, and Solaris, as well as NAS servers from QNAP and Synology.

Rclone v1.48 is the current version. Your distribution may use an earlier version (e.g., Debian "Stable" uses version 1.35). However, you can install the latest version of Rclone directly with:

curl https://rclone.org/install.sh | sudo bash

You may have to install cURL first by typing:

sudo apt install curl

Creating Remotes

You control Rclone with the command of the same name (rclone) and an attached command line including various options (see Table 1). To get started, call the configuration with rclone config. The setup wizard offers a number of options: Use S to set a password and Q to quit the program.

Table 1

Important Rclone Commands

Command

Function

rclone config

Opens the configuration

rclone copy

Copies data from the local system to cloud storage

rclone sync

Synchronizes source and target

rclone mount

Mount the drive locally into the filesystem

rclone move

Moves data from source to destination

rclone delete

Removes the contents of a path

rclone purge

Removes the path and its contents

rclone ls

Lists all objects in the path with a size specification

rclone md5sum

Creates md5sum checksums for all objects in the path

rclone size

Number and total size of all objects in <Remote>:<Path>

rclone cryptcheck

Checks the integrity of an encrypted remote

These and other commands show their options after entering -h.

Then use N to set up the first remote connection ("remote" here refers to the connection to a specific folder on a specific cloud service) – select this option first. Since Rsync lets you configure multiple remotes, make sure you use a meaningful name in the first step. Then select the desired service or protocol using a number (Figure 1).

Figure 1: In the interactive configuration dialog, create a name for the remote access and select the desired service or protocol from the list.

Google Drive Tested

In my hands-on test, I set up Rclone with Google Drive as a proprietary provider on a laptop. Of course, I do not want to store private data on Google's servers without encryption. Therefore, after an initial open test, I encrypted the data to be backed up before uploading with Rclone.

In the test, the remote repository is named gdrv. You then need to select the corresponding number for the desired service (Google Drive here) from the list by entering 12. For each provider and each protocol, the documentation offers a detailed manual [3], including Google Drive. The following two queries after client_id and client_secret can be skipped by pressing Enter.

Then decide whether Rclone should have full access to all files of the online repository or only access its own uploaded data. In addition, access can be restricted so that Rclone can only read. I selected full access for the test with the 1 option. Then skip the question about the root_folder_id, as well as the question about the service_account_file – Rclone can determine the root itself.

Can It Do That?

Finally, it doesn't hurt to look at the advanced configuration for Google Drive to learn about the additional options detailed in the documentation [4].

You are then taken to the point where Rclone can access your Google Drive account. To do this, say Y to the auto config prompt, and open your Google login in a browser window. After logging in, confirm that Rclone is allowed to access your drive (Figure 2). In the terminal, a Got Code message will appear; you can now close the browser again.

Figure 2: The dialog gives Rclone access to your Google Drive instance. If the connection fails to open, entering the link manually can help.

Having said this, the automatic method did not work for me. It was possible to register with Google, but the Rclone connection did not happen. A second attempt where I manually entered the URL, http://127.0.0.1:53682/auth, was then successful and let me complete this remote's configuration. To run Rclone in headless mode (i.e., without the option of opening a browser), create an API key in JSON format via a Google Service Account [5] and enter it in the advanced configuration.

This completes the setup wizard. You now have the choice of configuring additional remotes or protecting the configuration with a password. To do this, restart the configuration and select S for "set configuration password". Instead of a password, you can also make the configuration file readable for only your user with:

chmod 600 ~/.config/rclone/rclone.conf

Buy this article as PDF

Express-Checkout as PDF
Price $2.95
(incl. VAT)

Buy Linux Magazine

SINGLE ISSUES
 
SUBSCRIPTIONS
 
TABLET & SMARTPHONE APPS
Get it on Google Play

US / Canada

Get it on Google Play

UK / Australia

Related content

  • Workspace – rclone

    As a simple and reliable command-line backup utility that supports popular cloud storage services, rclone makes a perfect tool for maintaining an off-site backup of your data. This article can help you to get started.

  • The sys admin's daily grind: Rclone

    Having a good backup is a matter of course for sys admin columnist Charly Kühnast, but devices could still fall victim to fire or theft some day. Because he has enough free space on Google Drive, he doesn't need to search long for a solution. The only thing missing is the right tool.

  • Google Drive File Sync with rclone
  • Website Backup

    Shared hosting is the best way for first-time webmasters to get started. But what do you do about backup?

  • FOSSPicks

    Graham nearly made it through an entire month of FOSSPicks without an esoteric audio discovery. And then he found VeeSeeVSTRack.

comments powered by Disqus

Direct Download

Read full article as PDF:

Price $2.95

News