What we do · Data and web
Database replication and off-site replicas

A continuously updated, read-only copy of your production database, running at a different site, where reports can run without touching production and where you can fail over if the primary site is ever lost. Built so the replica reaches in to fetch changes, and production never has a way out to it.
What it is
Asynchronous MySQL or MariaDB replication, with the replica placed somewhere a different failure cannot reach: another datacenter, another storage array, another network path. The replica follows the primary in near real time, and it serves three jobs at once:
- A disaster recovery copy of the data, current to seconds.
- A reporting and integration target, so heavy read jobs stop competing with your customers.
- A clean source for backups, taken without locking or slowing the primary.
How we build it
The replica pulls. Replication connections are opened by the replica, never the other way round. Production has no network path into the replica side, enforced on the firewall appliances, not by a host rule someone can switch off. If a production server is ever compromised, the attacker has no route to the copy you would recover from.
Over our private mesh. Off site replicas travel over our encrypted private backbone between our datacenters in Athens, Amsterdam and Toronto, never across the public internet in the clear.
GTID based. We use global transaction identifiers, so a replica can be re-pointed, promoted or reseeded without anyone hunting for binary log coordinates at three in the morning.
Monitored on state, not noise. A check runs every five minutes and posts one alert when the state changes: lagging beyond the threshold, broken, or recovered. Engineers get a signal they act on, not a stream they learn to ignore.
What it protects you from
- Losing a whole site and with it the only copy of your data.
- Reporting queries and integrations slowing the database your customers are using.
- A replica that silently stopped weeks ago, discovered only on the day you need it.
- An attacker on a production server reaching the copy you would restore from.
- Backups that lock tables or stall the primary while they run.
In practice
A broadcaster's move off VMware. When we rebuilt a national broadcaster's web platform on our OLVM private cloud, the database was the only tier carrying state worth moving. It travelled by GTID replication from the old platform to the new one, so the cutover itself was a promotion with seconds of delta. The cutover completed months ago and the platform has run on our side ever since.
Replicas that serve new applications. A read replica is also the cleanest way to put a new search, reporting or portal front end on top of an existing system: the new application reads at local network speed and never touches the primary during page loads. We build these replica initiated, so the host carrying the replica needs no inbound firewall rule.
Why EuroVPS
- Replicas live at a genuinely different site, joined by our own encrypted backbone.
- Pull only, enforced by the firewall: production can never reach the replica side.
- State change alerting every five minutes, read by engineers around the clock.
- GTID throughout, so promotion and reseeding are rehearsed procedures.
- The same team builds the replica, the backups and the failover plan, so they fit together.
FAQ
How far behind will the replica be? Under normal load, seconds or less. We alert when lag crosses a threshold we agree with you, and the RPO you get follows the service you order.
Can we fail over to the replica? Yes. Promotion is a documented procedure. How quickly applications follow depends on the service design, from a manual promotion for a basic service to an automated, rehearsed switchover for globally covered ones.
Can our reporting tool connect to it? Yes, with a read only account. That is one of the main reasons to have it.
Does it replace backups? No. A replica copies mistakes as faithfully as data. Backups are taken alongside it and pulled to separate infrastructure.
MySQL or MariaDB? Both. We match the replica's engine and settings to the primary.
Want a copy of your data that a bad day cannot reach? Talk to an engineer.
Your next infrastructure partner
Let’s look at
your setup.
Tell us what you run today and what needs to improve. An engineer will review your brief and help you plan the next step.
+357 222 32280
who will look after your infrastructure.
Let’s start with your setup.
A little context helps us start the right conversation.
