Bash script to crate backup archives of Logseq graphs, inspired by the workflow of Joplin Backup Plugin. Tested on It can run automatic backups through Systemd.
This script is designed to create a compressed and optionally encrypted backup of a Logseq graph directory using 7zip, so you can safely store them also on removable or cloud storage. If a password is provided by the user, both file names and contents are encrypted with AES-256 encryption.
- Options such as graph directory and backup paths can be given on the command line or, better, in the
~/.config/logseq-backup.confconfiguration file (use the--create-confoption to create a template config file to customise). - Backup archives are named after current date and local hostname using this pattern, so you can tidily keep archives from multiple devices on the same location:
logseq-backup-YYYY-MM-DD_HH.MM.SS.hostname.7z - If no changes from last backup are detected, it doesn't create any new backup archive.
- The user can provide a maximum number of backups to keep, to automatically remove older backups.
- This script can create Systemd unit files to run backups periodically and every time the user logs in.
Requires 7z binary to create archives. You can get it running sudo apt install p7zip* on Debian/Ubuntu, sudo dnf install p7zip* on Fedora or brew install p7zip on macOS.
To enable (optional) automatic backup you need a Systemd-based distro (sorry macOS users!).
Then download, fork or copy and paste the script to your machine, put it in ~/.local/bin or somewhere in your path and make it executable.
$ cp logseq-backup.sh ~/.local/bin/
$ chmod +x ~/.local/bin/logseq-backup.shAutomatic backups are done via systemd. To install the required systemd unit files:
$ logseq-backup.sh -iIf not needed any more, you can uninstall them with:
$ logseq-backup.sh -u$ logseq-backup.sh --help
Usage: logseq-backup.sh [OPTION]
Create logseq graph backups. Use parameters from command line or from config file
in ~/.config/logseq-backup.conf
Command line parameters override config file ones
-n NOTEPATH Logseq graph path directory
-b BACKUPPATH Backup directory path
-p PASSWORD Encryption password
-f FILENAME Backup archive file name
-c Create a template config file in ~/.config/logseq-backup.conf
-i Setup systemd unit files to automate backups
-u Remove systemd unit files and disable automatic backups
-h Display this helpLet's say that your notes are in ~/Documents/logseq and you want to backup them to ~/Backup/notes/, encrypting them with the password foobar:
$ ./logseq-backup.sh -n ~/Documents/logseq/ -b ~/Backup/notes/ -p foobar
Changes detected: let's create a new backup
Changes detected: Creating a new backup archive...
7-Zip [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=it_IT.UTF-8,Utf16=on,HugeFiles=on,64 bits,12 CPUs Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz (906EA),ASM,AES-NI)
Scanning the drive:
7 folders, 22 files, 28151425 bytes (27 MiB)
Creating archive: /home/user/Backup/notes//logseq-backup-2024-03-11_19.27.23.carbon.7z
Items to compress: 29
Files read from disk: 22
Archive size: 20426585 bytes (20 MiB)
Everything is Ok
Backup of /home/user/Documents/logseq/ in /home/user/Backup/notes//logseq-backup-2024-03-11_19.27.23.carbon.7z completed.
Looking form excess archives to remove...
No excess backups to remove.Let's say we want to automate previous backup, running it at every login and twice a day.
First, put the script in your path:
$ mv logseq-backup.sh ~/.local/binThen create a template configuration file:
logseq-backup -cEdit the configuration file in ~/.config/logseq-backup.conf as needed. For example:
# Logseq graph dir to backup
note_dir=~/Documents/logseq
# Directory to save archives to
backup_dir=~/Backup/notes/
# Encryption password
password=foobar
# How often should backups run?
backup_interval=12h
Install systemd unit files to automate backups:
$ logseq-backup.sh -iThis way, at next login and 12 hours later you'll get a new package, if there are changes in your Logseq graph.
Backup are useless if you can't restore them. From time to time you should try to restore some of your backup archives to be sure everything went fine.
Luckily, on Logseq you can open a second (or third, ...) graph alongside with your current one.
First unpack your archive (you should also be able to unpack it from the file manager using file-roller):
$ 7z x logseq-backup-2024-03-05_09.58.49.carbon.7z
7-Zip [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=it_IT.UTF-8,Utf16=on,HugeFiles=on,64 bits,12 CPUs Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz (906EA),ASM,AES-NI)
Scanning the drive for archives:
1 file, 5884202 bytes (5747 KiB)
Extracting archive: logseq-backup-2024-03-05_09.58.49.carbon.7z
Enter password (will not be echoed):
--
Path = logseq-backup-2024-03-05_09.58.49.carbon.7z
Type = 7z
Physical Size = 5884202
Headers Size = 4458
Method = LZMA2:23 7zAES
Solid = +
Blocks = 1
Everything is Ok
Folders: 38
Files: 198
Size: 6985507
Compressed: 5884202Then start Logseq, open left sidebar, click on logseq > Add new graph and point it to the directory you just restored:
And check if everything is ok. You can switch on and off from your actual graph and restored graph using the same menu.
If you don't need it any more, you can safely remove it and unlink it visiting the All graphs page from the logseq menu.
I love Logseq, and use Syncthng to sync my note graph between my devices. One day I got several sync conflicts between my desktop and laptop (it was my fault), and after spending too much time trying to resolve them, I decided I needed a way to take snapshot of my graph I can keep as a reference or as a quick restore point in case of disaster.
I used Joplin in the past, and liked the workflow of the Backup plugin, so I wrote this script, trying to replicate that workflow.