User guide

Documentation

From nothing installed to a scheduled, verified, off-site backup. About fifteen minutes, most of it waiting for the first run.

The short version. Install → New backup job → connect to your server → tick the databases → add a destination → set a schedule → Run now once and watch it finish. Everything after that is refinement.

1. Install

Download the installer from the download page and run it. Windows asks for administrator approval — it is registering a service and creating a data folder outside Program Files. There is nothing to install first; the .NET runtime is inside the package.

Then sign out of Windows and back in. This is the step people skip, and it is the difference between the console working and appearing broken. The installer creates a LeoCore Operators group and adds you to it, but Windows stamps group membership into your logon token when you sign in and never refreshes it. Until you sign out and back in, your session cannot see the new group and the console will say your account is not allowed to manage LeoCore Backup. Rebooting works too. In a hurry, right-click the console and choose Run as administrator.

Full detail, including unattended and Server Core installs, is on the installation page.

2. Create your first job

A job is one database server, the databases you picked on it, where copies go, and when. Most sites need one job per server.

Click New backup job in the sidebar. The connect dialog opens straight away. Rename the job later with the pencil icon beside its name.

3. Connect to the database server

Pick your engine, then fill in the connection.

SQL Server	Leave the server as `localhost` for an instance on this machine, or use `HOST\INSTANCE`. If instances are found on this machine they appear as chips — click one to fill the box. Windows Authentication uses the backup service's account.
MySQL / MariaDB	Host and port 3306. LeoCore finds `mysqldump` automatically; if it is somewhere unusual, set the bin folder under Advanced settings.
PostgreSQL	Host and port 5432, plus the `bin` folder containing `pg_dump` if it is not on the standard path.
Firebird	Host and port 3050 with the SYSDBA account. Firebird cannot list its own databases, so you add each one by file path — see below.
Azure SQL / Amazon RDS	The endpoint hostname and a SQL login. These are managed services, so LeoCore uses the engine's own export rather than a file-level backup.

Click Connect & list databases. If it fails, the message says which part failed — the network, the credentials or the permissions — rather than just "could not connect".

4. Choose the databases

The right-hand pane lists what was found, with sizes and recovery models. Tick the ones you want. Select all / none is at the bottom, and the filter box helps on a server with two hundred databases.

Firebird works differently, and it is not a limitation of LeoCore. A Firebird database is a file, and the server only knows about one once something connects to it — there is no way to ask a Firebird server what databases exist. So you type the path (C:\Data\ACME.FDB) or an alias from databases.conf and click Add. On the server itself, Browse… opens a file picker. The path is where the database sits on the server, not on the machine you are sitting at.

5. Choose where the copies go

On the job dashboard, Destinations → + Add. Every run uploads to every destination you list, so two destinations means two independent copies.

Local folder, network share, FTP/SFTP, Amazon S3, S3-compatible, Azure Blob, Google Drive, Dropbox, OneDrive, OneDrive for Business, Box, Backblaze B2 and Yandex.Disk.

At least one copy should leave the machine. A backup on the same disk as the database protects you from a bad UPDATE, and from nothing else — not a failed drive, not ransomware, not a fire. If you take one thing from this guide, take this one.

Where the destination supports it, turn on immutability (S3 Object Lock, Azure immutable blobs). An immutable copy cannot be deleted or altered before its retention expires, including by someone using your own credentials — which is exactly what ransomware does.

6. Set the schedule

Schedule → Edit. Three independent cadences:

Full	A complete copy. On its own it is enough for most sites.
Differential	Everything changed since the last full. Faster and smaller; needs its full to restore.
Transaction log	The change record since the last log backup. This is what makes point-in-time recovery possible — restoring to 14:32, not just to last night.

A common shape: full nightly, differential every six hours, log every fifteen minutes. If you are unsure, start with a nightly full and add the rest once that is proven.

The dialog shows the next several runs as you change the settings, so you can see what you have actually asked for before saving.

7. Decide what to keep

Retention is per job. Keep the last N backups, or use grandfather-father-son to hold daily, weekly, monthly and yearly copies at different depths.

LeoCore will not delete a backup another one still needs. A differential is useless without its full, and a log chain is useless with a gap in it. Retention rules are applied to whole restorable chains, so a rule that would strand a backup keeps the parent instead of quietly making both worthless.

8. Run it once, on purpose

Click Run now and watch it. The log streams live: dump, compress, encrypt, upload, verify. It is worth watching the first one all the way through — permissions and destination credentials fail here, where you are looking, rather than at 2am.

Archives are compressed and, if you leave the default on, encrypted with AES-256. The archive password is generated once and stored on the machine.

Export the archive password and keep it somewhere else. Settings → Export the archive password. Encrypted archives cannot be restored without it. If the machine you are backing up is also the only place that password exists, then a fire takes both — and your off-site copies become unreadable files.

9. Prove it restores

Verification & restore drill → Configure. On a schedule you choose, LeoCore restores the newest backup into a scratch database, runs a validation query, records the result and throws the scratch copy away.

A backup that has never been restored is a hypothesis. This is the feature that turns it into a fact, and it is the one most worth turning on.

10. Get told when something breaks

Email notifications → Configure. Enter recipients and choose whether to send on success, on failure, or both. Reports go over HTTPS through our hosted relay — nothing to configure, no SMTP server needed, and no port 25 on a database host.

If you would rather use your own mail server, Settings → Email takes SMTP details. Your SMTP server replaces our relay entirely for every job, and you choose the From address.

Most sites should send on failure only. A success email every night becomes wallpaper, and then nobody notices the night it does not arrive.

11. Restoring

Two routes, both from the job dashboard:

Recent runs → Restore — restore that specific backup.
Point-in-time recovery → Open — pick any instant the chain covers. LeoCore works out which full, differential and log backups to replay, and tells you before you start if a link is missing.

You can restore over the original database, or as a new database with a different name — which is what you want when checking what a table looked like last Tuesday without touching production. Restoring to a different server is supported too.

Practise once before you need it. Restore as a new database on a test machine while nothing is wrong. Ten quiet minutes now is worth a great deal at the point where you are doing it for real with people waiting.

12. Entering your licence

LeoCore runs as the free Solo edition until you enter a key. Settings → Licence, paste the key from your purchase email, click Activate.

For a server with no internet access, use offline activation: the console shows a request code, you paste it into that page from any machine, and paste the response back.

Keep the email. The key is how you move the licence to replacement hardware. If you have lost it, write to admin@seksolution.com from the address you bought with and we will resend it.

Stuck on something this page does not cover? The FAQ answers the questions we are actually asked, and support@seksolution.com replies within one business day.