Backup & Restore

Daily automatic backup + quick recovery is the two legs of data safety.

Where data lives

Type	Path	Importance
SQLite main DB	./data/one-api.db (container /app/data/)	Critical
MySQL data volume	./mysql-data/	Critical
Redis data volume	./redis-data/	Important (inter-node state)
User uploads	./data/uploads/	Depends on usage
Exported logs	./data/logs/	Low

Recommended strategy

1. Daily database full backup (required)

SQLite:

docker exec one-api-pro \
  sqlite3 /app/data/one-api.db ".backup '/app/data/backup-$(date +%Y%m%d).db'"

# or stop-then-copy:
docker stop one-api-pro
cp ./data/one-api.db ./backup/one-api-$(date +%Y%m%d).db
docker start one-api-pro

MySQL:

docker exec one-api-pro-mysql \
  mysqldump -uoneapi -poneapi-pass oneapi | gzip > ./backup/mysql-$(date +%Y%m%d).sql.gz

Automate: add to crontab, run at midnight, retain 30 days.

2. Weekly uploads backup (if used)

tar czf ./backup/uploads-$(date +%Y%m%d).tgz ./data/uploads/

3. Off-site backup (required in production)

Sync ./backup/ to object storage:

• AWS S3 / Aliyun OSS / Tencent COS
• rsync to another machine
• 7 days → 30 days → 1 year tiered

Tools: restic / borgbackup / cloud-provider CLIs.

Restore

# 1. Stop
docker compose down

# 2. Replace DB
docker compose up -d mysql
gunzip < ./backup/mysql-YYYYMMDD.sql.gz | \
  docker exec -i one-api-pro-mysql mysql -uoneapi -poneapi-pass oneapi

# 3. Restart all
docker compose up -d

Verify after restore

• [ ] Log in to the admin console
• [ ] Admin → Users — count matches
• [ ] Admin → Channels — configurations intact
• [ ] Admin → Plans — plans still active
• [ ] Send a test call — log records it

Related

• Docker Single-instance
• Docker Compose
• System Requirements