Skip to content

Restore a server backup

Download a server backup as one archive or in parts, then restore files and MySQL / MariaDB databases from it on the command line.

Last updated

Server backups are restored by you, on the server or wherever you need the files: you download the archive from the panel and extract it. WP Safer does not write anything back to your server — there is no restore button that overwrites files or databases for you. This article shows how to download a backup and restore files and MySQL / MariaDB databases from it.

Download the backup

  1. Open the server and go to the Backups tab.
  2. In Backup history, find the backup you need. Completed and partly completed backups have a Download button.
  3. The Download backup window offers Download the whole archive as one .tar.gz file. Large backups are stored in parts of 64 MB; the window also lists each part separately, which lets you resume an interrupted download.
  4. These links work for a limited time and only for you. Open the window again for fresh links.

A database backup can also be downloaded one database at a time from the Databases tab, without the rest of the archive.

If you downloaded parts, join them first

The parts are consecutive pieces of the same archive. Join them in order before extracting; the window shows the exact command for that backup, for example:

cat web-1-job3-20261031-020000.tar.gz.* > web-1-job3-20261031-020000.tar.gz

Restore files

Files are stored relative to /, so /var/www/site/index.php is in the archive as var/www/site/index.php. Extract into a folder of your choice first and copy back what you need:

mkdir -p /restore
tar -xzf web-1-job3-20261031-020000.tar.gz -C /restore

Extracting straight into / overwrites the live files with the backed-up ones, including ones that changed since the backup. Only do that on purpose.

Restore a MySQL / MariaDB database

Each database is in db/<name>.sql. Extract the archive, then import a database:

tar -xzf web-1-job4-20261031-030000.tar.gz db/
mysql < db/<name>.sql

Very large dumps are stored in numbered pieces (db/<name>.sql.0001, .0002 …). Join them with cat in order, or import them in one go:

cat db/<name>.sql.* | mysql

The dump recreates the database and its tables. To restore into a different database name, import with mysql <other-name> < db/<name>.sql after removing the CREATE DATABASE and USE lines at the top of the file.

Check the archive before you rely on it

  • Every run is checked before it is marked completed: the archive must be a complete gzip stream and a complete tar archive, every part must arrive with the checksum it was sent with, and each database dump must end with -- Dump completed.
  • A Partly completed run is missing at least one item (a folder that could not be read or a database whose dump failed). The rest of it can still be restored.
  • Test a restore now and then on a spare machine. A backup you have never restored is a backup you hope works.