WebServerCloudBackups

Automatic backups your web projects bases (MySQL/MariaDB) and files to the clouds via WebDAV or Amazon S3 and to backup servers via SSH (rsync). Supports setting passwords for archives (WebDav/S3) and excluding specified folders.

Requirements

• curl (for WebDAV)
• s3cmd (for S3)
• 7zip archiver (usually p7zip-rar p7zip-full on deb-based distros)
• connection to backup server via SSH key without passphrase (for ssh)

Configuring

1. Login server in root user

2. Copy backup.sh and backup.conf to other directory on your server.

3. Change permissions of backup.conf to 600.

4. Declare main vars in backup.conf

• MYSQL_USER - MySQL/MariaDB user (min user privileges: EVENT,LOCK TABLES,SELECT,SHOW DATABASES on all databases);
• MYSQL_PASS - MySQL/MariaDB user password;
• CLOUD_USER - login for your cloud (for WebDAV);
• CLOUD_PASS - password for your cloud user (for WebDAV);
• CLOUD_PATH - full path to cloud folder for WebDAV (ex.: https://webdav.yandex.ru/Backups/) or path to S3 spacename (ex.: s3://myspacename);
• CLOUD_PROTO - cloud protocol (webdav or s3 or ssh, default value is webdav if empty or undefined);
• CLOUD_SSH_HOST - hostname/IP and port of backup server separated by colon (for ssh; ex.: 123.123.123.123, 123.123.123.123:2222, hostname.com:4444);
• CLOUD_SSH_HOST_USER - system username of backup server (for ssh);
• CLOUD_SSH_HOST_PATH - full path to backups dir on backups server (for ssh, projects dirs will be created automatically) ;
• TMP_PATH - path for temporary files on server (ex.: /tmp/);
• GLOBAL_ARCHIVE_PASS - global password for created archives (if project password set to false it will be used this password. if project password set to false and this password set to false password not set to project archive.);
• EXCLUDE - spaces separated folders to exclude (supports wildcard in folders names, ex.: EXCLUDE=".svn .git *cache*");
• EXCLUDE_RELATIVE - relative folders paths to exclude separated by spaces (supports wildcard in paths to folders, ex.: EXCLUDE_RELATIVE="wp-content/cache templates/*_temp");
• <PERIOD>_EXCLUDE - spaces separated folders to exclude for specific backup period (ex.: DAILY_EXCLUDE="uploads");
• <PERIOD>_EXCLUDE_RELATIVE - relative folders paths to exclude separated by spaces for specific backup period (ex.: WEEKLY_EXCLUDE_RELATIVE="wp-content/uploads");
• SPLIT - size of archive parts (set false if you don't want split archives into parts); supports b (bytes), k (kilobytes) m (megabytes) g (gigabytes) (ex.: SPLIT="500m");
• LAST_BACKUPS_PATH - folder for lists of last backup files (script use its for deleting old files from cloud to avoid errors and unnecessary files with splitting archives into parts; folder create automatically; this folder is in the same folder as the main script with name last_backups if this var not set);
• SCRIPT_LOG_PATH - full path for logs (directory must be exists and current user must be have permissions for write into it; logging disabled if path not setted);
• SORT_BACKUPS (true|false) - sort backups by subdirectories in cloud (files for files, databases for databases; disabled by default).

Note: relative and not relative lists will be united if using ssh proto.

5. Add your projects after declare -A projects one per row like below:

projects[unique_key]="<project_name> <db_name> <project_folder> <project_archive_password>"

Parameters in quotes must be written through spaces and all required.

Parameters:

• <project_name> - project name, you must create folder with same name in the cloud folder, defined in CLOUD_PATH
• <db_name> - database name, type false if database backup is not required for project
• <project_folder> - full path to project folder, type false if files backup is not required for project
• <project_archive_password> - project archive password, type false if password is not required for project archive or using global password, defined in GLOBAL_ARCHIVE_PASS

You can specify backup method to project or to files or database separatelly. Just add protocol to project name or database name or files path via colon.

NOTE: you can't use WebDav and S3 protocols together (only WebDav+SSH or S3+SSH). Ability to use all protocols at the same time will be added later.

Examples

Use default protocol defined in CLOUD_PROTO var:

projects[1]="domain.org false /home/user/www/domain.org false"
projects[2]="domain.com com_db /home/user/www/domain.com 1234"

Do backup all project via ssh:

...
CLOUD_PROTO="webdav"
...
projects[1]="domain.com:ssh com_db /home/user/www/domain.com false"

Do backup project database via WebDav and files via default proto (ssh):

...
CLOUD_PROTO="ssh"
...
projects[1]="domain.com com_db:webdav /home/user/www/domain.com false"

Do backup project database via default proto (WebDav) and files via ssh:

...
CLOUD_PROTO="webdav"
...
projects[1]="domain.com com_db /home/user/www/domain.com:ssh false"

Usage

Directly in shell

bash backup.sh <backup_type> <period> <compress_ratio> <enabled_protocol>

Compression ratio parameter is opional. It sets to 5 if it not set. It required if you want set another protocol (for backward compatibility).

Examples:

bash backup.sh bases daily
bash backup.sh files weekly 7
bash backup.sh files weekly 5 ssh

Supported backup types:

• files - backup projects folder
• bases - backup projects bases

Supported periods:

• hourly - add "hourly" mark and current hour to archive name (ex.: domain.com_base_hourly_02.7z)
• daily - add number and name of the current week day to archive name (ex.: domain.com_files_5_Friday.7z)
• daily_week - add number of the current week day to archive name (ex.: domain.com_base_daily_week_2.7z)
• daily_month - add number of the current month day to archive name (ex.: domain.com_base_daily_month_05.7z)
• weekly - add "weekly" mark to archive name (ex.: domain.com_files_weekly.7z)
• monthly - add "monthly" mark to archive name (ex.: domain.com_files_monthly.7z)

Supported compress ratio:

• 0 - without compression
• 1 - fastest
• 3 - fast
• 5 - normal (default)
• 7 - maximum
• 9 - ultra

Note: better compression ratios with big files can lead to fails. if at an archiving there is a fails that it is necessary to lower compression ratio.

Supported protocols:

• webdav - WebDav
• s3 - Amazon S3
• ssh - via rsync (required connection to backup server via ssh key)

Note: by default the script will do backups with all protocols. The script backup only with specified protocol if it set.

Cron

Add lines in root crontab like below

10 * * * * /bin/bash /path/to/script/backup.sh bases hourly # bases backup every hour in 10 minutes
0 0 * * * /bin/bash /path/to/script/backup.sh bases daily # bases backup every day in 00:00
20 0 * * 1 /bin/bash /path/to/script/backup.sh bases weekly # bases backup every monday in 00:20
40 0 1 * * /bin/bash /path/to/script/backup.sh bases monthly # bases backup every 1st day every month in 00:40
0 1 * * 1 /bin/bash /path/to/script/backup.sh files weekly 7 # files backup every monday in 01:00 with changed compression ratio
0 4 1 * * /bin/bash /path/to/script/backup.sh files monthly 7 # files backup every 1st day every month in 04:00 with changed compression ratio
0 1 * * 1 /bin/bash /path/to/script/backup.sh files weekly 5 ssh # only files backup via rsync every monday in 01:00

If you want receive script result to email add below to the top of crontab list (require working MTA on your server)

MAILTO=name@domain.com

Tested on

• Hetzner Storage Box (WebDav, SSH)
• Yandex Disk (WebDav, not recommended for big files)
• Mail.ru Cloud (WebDav)
• DigitalOcean Spaces (S3)
• backup servers (SSH)

TODO

• add support for others database types backup
• add support for partitioning archives into specified size
• add automatically checking/creating folders in cloud
• add logging
• validating vars from config file
• add full support for some special characters, spaces and non latin characters in file names and paths
• add ability to backup files and databases to own archive
• add functionality for restore from backups
• add support for local backup to mounted clouds disks
• add support for backups via rsync
• make package with system daemon and flexible backups customization
• refactor this sh*t

Changelog

• 23.06.2024 - 1.10.0 - Added support for daily backups for each day of month
• 23.03.2024 - 1.9.2 - Added --ignore-missing-args to rsync for suppress vanished files warnings
• 18.03.2024 - 1.9.1 - Fixed files exclusion with ssh protocol
• 17.02.2024 - 1.9.0 - Added support for sorting backups by subdirectories
• 27.01.2024 - 1.8.1 - Bugfix with enabled --single-transaction mode
• 21.01.2024 - 1.8.0 - added logging, escaping values of sensitive vars, refactoring, bugfixes
• 21.07.2023 - 1.7.0 - added support for backups via rsync
• 14.03.2022 - 1.6.2 - added new parameters to mysqldump command + gzip compression
• 17.12.2020 - 1.6.1 - fixed archive filename for hourly backup period
• 13.12.2020 - 1.6.0 - added hourly backup period
• 29.07.2020 - 1.5.0 - added parameters for excluding folders for specific backup periods
• 17.04.2020 - 1.4.1 - removed "--databases" parameter in mysqldump command for support restore databases to another databases
• 23.11.2019 - 1.4.0 - added support for S3 storages
• 24.10.2019 - 1.3.2 - bug fixes
• 23.10.2019 - 1.3.1 - hidden curl success messages for clean output
• 18.02.2019 - 1.3.0 - added support for sets user for databases backups instead using root user
• 01.03.2018 - 1.2.1 - bug fixes
• 22.02.2018 - 1.2.0 - added support for partitioning archives into specified size, automatically checking/creating folders in cloud, automatically remove slashes in the end of paths in vars to avoid errors, small code refactoring
• 20.02.2018 - 1.1.0 - added support for excluding folders when archiving, small code refactoring, added new vars to config file
• 14.05.2017 - 1.0.2 - added compress ratio parameter
• 13.05.2017 - 1.0.1 - main script code refactoring
• 11.05.2017 - 1.0.0 - released