Backup to S3 — online backup in Uzbekistan
A server fails, a disk dies, a developer drops the database by accident — without backups, recovery drags on for days or becomes impossible. With backups in S3 object storage it's a matter of minutes: copies live separately from the server, are available around the clock over HTTPS, and capacity grows automatically.
Why S3 fits backups
Keeping backups on the same server is a common mistake: a hardware failure or a breach loses both the data and the copies. Object storage solves this:
- Independent of the server. If the server fails or data is deleted, the copies in S3 stay intact.
- Scalable. Storage grows with the project and you pay for actual usage — no need to pre-allocate disks.
- Available 24/7. You can restore from any server at any time over HTTPS.
- Tool compatibility. Works with rclone, restic, duplicati, AWS CLI and any S3-compatible software — no custom integration.
- Data in Uzbekistan. SERVERPLUS storage sits in a local datacenter connected to TAS-IX: low latency and compliance with data-localization requirements.
What you can store
- Databases — MySQL, PostgreSQL, MS SQL, MongoDB, Redis.
- Website and web-app files.
- Server and application configs.
- Media and user data.
- Logs and archives.
- Docker images and build artifacts.
- Virtual machine snapshots.
How it works
The scheme is simple: you create a dedicated bucket for backups and set up a tool (rclone, restic or the DBMS's own utilities) that uploads copies to the cloud on a schedule. Restore is the reverse: download the needed file or dump and deploy it on the server.
Quick start
The storage endpoint (S3-compatible):
https://object-store.serverplus.uzCreate a bucket and keys in the cloud.serverplus.uz panel ("Object storage" → "Create bucket"; "Access" → "S3/EC2 credentials"), then connect a tool.
Tools and examples
rclone — universal sync. Config ~/.config/rclone/rclone.conf:
[sp]
type = s3
provider = Other
access_key_id = <ACCESS_KEY>
secret_access_key = <SECRET_KEY>
endpoint = https://object-store.serverplus.uzrclone sync /var/www sp:backups/www # site files
rclone copy dump.sql.gz sp:backups/db/ # database dumprestic — incremental encrypted backups with deduplication:
export AWS_ACCESS_KEY_ID=<ACCESS_KEY>
export AWS_SECRET_ACCESS_KEY=<SECRET_KEY>
export RESTIC_PASSWORD=<encryption-password>
restic -r s3:https://object-store.serverplus.uz/backups init
restic -r s3:https://object-store.serverplus.uz/backups backup /var/www /etcDB dumps + AWS CLI:
pg_dump mydb | gzip | aws --endpoint-url https://object-store.serverplus.uz \
s3 cp - s3://backups/db/mydb-$(date +%F).sql.gzduplicati — if you prefer a GUI: encryption, scheduling and versioning are configured in the browser.
Recommendations
- Rotation. Keep daily copies for 7 days and weekly ones for a month; set a lifecycle rule on the bucket to auto-delete old ones.
- A separate bucket for backups. Don't mix them with working files — easier to manage access and size.
- Test restores. Once a month: a backup that won't restore is useless.
- Encrypt sensitive data. restic encrypts the whole repository; encrypt dumps with personal data before uploading.
- Monitor success. Alert on a failed backup — a silent failure is only noticed when you already need the data.
Object Storage SERVERPLUS
SERVERPLUS object storage is located in an Uzbekistan datacenter connected to TAS-IX: low latency when uploading from local servers, compliance with data-storage law, and support in Russian. The first backup takes 15–20 minutes to set up.
Related: "S3 object storage on serverplus" (tools and API in detail) and "Backup strategy" (snapshots, volumes and S3 together).