Take a logical backup (dump)

Oracle

expdp system/password DIRECTORY=dpump_dir \
  SCHEMAS=hr DUMPFILE=hr_%U.dmp LOGFILE=hr_exp.log PARALLEL=4

Uses Data Pump. The DIRECTORY object must be defined beforehand with CREATE DIRECTORY. Use RMAN for physical backups.

PostgreSQL

pg_dump -U postgres -h localhost -Fc -f db.dump mydb
# -Fc = compressed custom format (restored with pg_restore)

-Fc allows flexible restores (selective tables/parallel). Use pg_dumpall for the whole cluster (including roles).

MySQL

mysqldump -u root -p --single-transaction \
  --routines --triggers --events mydb > db.sql

--single-transaction gives a lock-free consistent backup on InnoDB. Parallel options for large data: mydumper/myloader or MySQL Shell util.dumpInstance()/dumpSchemas() (mysqlpump was deprecated in 8.0.34 and removed in 8.4).

SQL Server

BACKUP DATABASE [MyDB]
TO DISK = N'C:\backup\MyDB_full.bak'
WITH COMPRESSION, CHECKSUM, INIT;

INIT overwrites the existing file; use NOINIT to append. For log backups use BACKUP LOG (in the FULL recovery model).

MongoDB

mongodump --uri="mongodb://localhost:27017" \
  --archive=db.archive.gz --gzip --db=mydb

On a sharded cluster, take it through mongos with the balancer stopped for consistency.

Redis

BGSAVE
# Takes an RDB snapshot in the background (dump.rdb).
# Non-blocking copy: redis-cli --rdb /backup/dump.rdb

BGSAVE runs in the background via fork; SAVE blocks the server (do not use in production). With AOF enabled, less data is lost.