Backup or disaster recovery: what a daily copy covers and what it doesn't

People mix the two up because they reach for both on the same bad day. A backup is an archive: the state of the site at a given time. Disaster recovery is the plan for bringing the site back after a serious failure: who does what, from which copy, on which server, and how fast.

A business can have flawless daily backups and still be offline for two days if nobody knows where the copies are, or if the only person with the panel password is on holiday.

Two numbers to set

RPO (Recovery Point Objective) is how much data you can afford to lose. With a copy made at 3 a.m., if the site fails at 5 p.m., you lose whatever changed in the last 14 hours: orders, comments, form submissions, saved posts.

RTO (Recovery Time Objective) is how long you can afford to be down. Count the whole path: until you notice the problem, until you find the right copy, until it's restored, and until you've checked the site works.

For a blog, an RPO of 24 hours is usually fine, since you rewrite a post or two. For a shop taking 50 orders a day, the same 24 hours can mean paid orders you no longer have any record of.

When one copy a day stops being enough

A daily copy handles the usual mistakes well: a plugin that breaks the theme, a page deleted by accident, an update that fails. It does have limits.

A shop takes orders all day. Restoring the database from last night's copy erases today's. Before you restore, export the new orders from WooCommerce so you can enter them again, or restore only the files if the problem is in the code.

An infected site can show no sign for weeks. If you keep only the last 7 copies, all of them may already contain the infected files, and then you need one from a month or two ago.

Copies kept on the same server as the site go down with the server. And a copy nobody has ever restored is an unknown: you don't know whether the archive is complete until you try.

How it works at AYT

On paid plans we copy the files and databases every night to a different server from the one the site runs on. We keep the last 7 days, one copy a week for the last 4 weeks, and the last copy of each month for 2 years. For a site infected three months ago, you still have a clean copy from before.

You restore it yourself from the panel, the whole site or just a few files, and you choose whether to bring back the files, the database or both. You can also download a copy of the site (.tar.gz for files, .sql.gz for the database) to keep on your side.

A one-page recovery plan

For a small or medium site, one page answering a few questions is enough.

1. Who notices the problem? Monitoring emails you when the site stops responding, but someone has to read that email on a Saturday too.
2. Who has access to the hosting panel, the domain registrar and DNS? At least two people, each with their own account and two-factor authentication.
3. Which copy do you restore from? For a problem that started today, last night's. For malware, the most recent copy you know for sure is clean.
4. What do you lose, and how do you get it back? That day's orders, form messages, new customer accounts.
5. Where do you keep a copy outside your host? A copy downloaded once a month to a disk or a cloud account of your own helps if you lose access to the hosting account, for example after an unpaid invoice or a compromised email.

Try a restore

Every few months, restore a copy somewhere other than the live site, for example on a test subdomain or on your own computer from the downloaded archive. Time it and write down what else you had to do afterwards: clear the cache, check the forms, reconnect the payment plugin. In a real emergency you'll already have the list.

Daily copies kept for 2 years on a separate server are included in every paid AYT plan.