Backups people actually use: what to back up, how often, and the restore drill nobody runs until they actually need it.
Every server owner says they have backups. Most of them have never actually tested a restore. The first time you find out, it's a Tuesday at 3am and your database is corrupted.
Three things: the MySQL database, the resources folder (especially your custom scripts), and the txAdmin profile / server.cfg. That's it. Everything else can be redownloaded.
Database: at least every 6 hours, ideally every hour during peak. Resources: daily, plus on every meaningful change. txAdmin / config: snapshot once a week and after any major change.
The cardinal rule: backups must live on a different machine than the server. A backup on the same VPS that died doesn't help anyone. Use cheap object storage (S3, Backblaze, Wasabi) or a second VPS in a different region.
Do this once a quarter, before you need it. Spin up a fresh test box. Pull a backup. Restore the database. Restart the server. Connect a test account. Verify your character loads. If any step fails, you don't have a backup — you have files in a folder.
mysqldump on a cron job, gzipped, uploaded via rclone to remote storage. A second cron job that prunes anything older than 30 days. A txAdmin recipe export weekly. Total setup time: about an hour. Ongoing maintenance: zero.
A working backup is one you have actually restored from. Until you have done a full restore drill at least once, all you have is hope.
Written by
Alex Johnson
Questions? Browse our products or contact us
