Backup

Everyone has backups. Few people have tested restores. Here’s how to build backup strategies that work when disaster strikes. The 3-2-1 Rule The foundation of backup strategy: 3 copies of your data 2 different storage types 1 copy offsite Example: Production database (primary) Local backup server (different disk) S3 in another region (offsite) This survives disk failure, server failure, and site failure. Database Backups PostgreSQL Logical backups (pg_dump): 1 2 3 4 5 6 7 8 # Full database pg_dump -Fc mydb > backup.dump # Specific tables pg_dump -Fc -t users -t orders mydb > partial.dump # Schema only pg_dump -Fc --schema-only mydb > schema.dump Physical backups (pg_basebackup): ...

Everyone knows backups are important. Few actually test them. Here’s how to build backup systems that work when you need them. The 3-2-1 Rule The classic foundation: 3 copies of your data 2 different storage types 1 offsite copy Example implementation: P C C C r o o o i p p p m y y y a r 1 2 3 y : : : : L R C P o e l r c m o o a o u d l t d u e c s b t n r a i a e c o p p k n s l u h i p d o c a t a ( t S a ( ( 3 b s d , a a i s m f d e e f i e f s r f e e e r n r v t e e n r d t , a t r d a e i g f c i f e o e n n r t ) e e n r t ) d i s k ) What to Back Up Always Back Up Databases — This is your business Configuration — Harder to recreate than you think Secrets — Encrypted, but backed up User uploads — Can’t regenerate these Maybe Back Up Application code — If not in Git, back it up Logs — For compliance, ship to log aggregator instead Build artifacts — Rebuild from source is often better Don’t Back Up Ephemeral data — Caches, temp files, sessions Derived data — Can regenerate from source Large static assets — Use CDN/object storage with its own durability Database Backups PostgreSQL 1 2 3 4 5 6 7 8 # Logical backup (SQL dump) pg_dump -Fc mydb > backup.dump # Restore pg_restore -d mydb backup.dump # All databases pg_dumpall > all_databases.sql For larger databases, use physical backups: ...

The only backup that matters is the one you can restore. Everything else is wishful thinking with storage costs. Here’s how to build backup systems that work when disaster strikes. The 3-2-1 Rule (Still Valid) The classic backup rule holds up: 3 copies of your data 2 different storage types 1 offsite location Modern interpretation: P L O r o f i c f m a s a l i r : t y e : : P D S r a 3 o i / d l G u y C c S t s / i n A o a z n p u s r d h e a o t t B a s l b o a ( b s d e i ( f d ( f i S e f S r f D e e ) n r t e n t o l r u e m g e i ) o n ) This protects against: ...

rsync is the standard for efficient file transfer. It only copies what changed, handles interruptions gracefully, and works over SSH. Here’s how to use it well. Basic Syntax 1 rsync [options] source destination The trailing slash matters: 1 2 rsync -av src/ dest/ # Contents of src into dest rsync -av src dest/ # Directory src into dest (creates dest/src/) Essential Options 1 2 3 4 5 6 -a, --archive # Archive mode (preserves permissions, timestamps, etc.) -v, --verbose # Show what's being transferred -z, --compress # Compress during transfer -P # Progress + partial (resume interrupted transfers) --delete # Remove files from dest that aren't in source -n, --dry-run # Show what would happen Common Patterns Local Backup 1 2 3 4 5 # Mirror directory rsync -av --delete /home/user/documents/ /backup/documents/ # Dry run first rsync -avn --delete /home/user/documents/ /backup/documents/ Remote Sync Over SSH 1 2 3 4 5 6 7 8 # Push to remote rsync -avz -e ssh /local/dir/ user@server:/remote/dir/ # Pull from remote rsync -avz -e ssh user@server:/remote/dir/ /local/dir/ # Custom SSH port rsync -avz -e "ssh -p 2222" /local/ user@server:/remote/ With Progress 1 2 3 4 5 # Single file progress rsync -avP largefile.zip server:/dest/ # Overall progress (rsync 3.1+) rsync -av --info=progress2 /source/ /dest/ Exclusions 1 2 3 4 5 6 7 8 # Exclude patterns rsync -av --exclude='*.log' --exclude='tmp/' /source/ /dest/ # Exclude from file rsync -av --exclude-from='exclude.txt' /source/ /dest/ # Include only certain files rsync -av --include='*.py' --exclude='*' /source/ /dest/ Example exclude file: ...

tar (tape archive) bundles files and directories into a single file. Combined with compression, it’s the standard way to package and distribute files on Unix systems. The Basics 1 2 3 4 5 6 7 8 # Create archive tar -cvf archive.tar /path/to/files # Extract archive tar -xvf archive.tar # List contents tar -tvf archive.tar Understanding the Flags 1 2 3 4 5 c = Create archive x = Extract archive t = List contents v = Verbose (show files) f = File (next arg is filename) So tar -cvf = Create, Verbose, File. ...

rsync synchronizes files between locations — local to local, local to remote, remote to local. It’s smart: it only transfers what’s changed, making it fast for incremental backups and deployments. Basic Syntax 1 rsync [options] source destination Local Sync 1 2 3 4 5 6 7 8 9 # Copy directory rsync -av /source/dir/ /dest/dir/ # Copy directory (trailing slash matters!) rsync -av /source/dir /dest/ # Creates /dest/dir/ rsync -av /source/dir/ /dest/ # Contents into /dest/ # Dry run (show what would happen) rsync -avn /source/ /dest/ Remote Sync (SSH) 1 2 3 4 5 6 7 8 9 10 11 # Local to remote rsync -av /local/dir/ user@remote:/remote/dir/ # Remote to local rsync -av user@remote:/remote/dir/ /local/dir/ # Different SSH port rsync -av -e 'ssh -p 2222' /local/ user@remote:/remote/ # With SSH key rsync -av -e 'ssh -i ~/.ssh/mykey' /local/ user@remote:/remote/ Common Options 1 2 3 4 5 6 7 8 9 -a, --archive Archive mode (preserves permissions, timestamps, etc.) -v, --verbose Verbose output -n, --dry-run Show what would be transferred -z, --compress Compress during transfer -P Progress + partial (resume interrupted transfers) --progress Show progress --delete Delete files in dest not in source -r, --recursive Recurse into directories -h, --human-readable Human-readable sizes Archive Mode (-a) -a is equivalent to -rlptgoD: ...

Everyone has backups. Few have tested restores. The backup that fails during a crisis is worse than no backup — it gave you false confidence. Backup strategy isn’t about the backup. It’s about the restore. The 3-2-1 Rule A minimum viable backup strategy: 3 copies of your data 2 different storage media/types 1 copy offsite P C C C r o o o i p p p m y y y a r 1 2 3 y : : : : P L R O r o e f o c m f d a o s u l t i c e t t s e i n r o a e a n p p r s l c d h i h a o c i t t a v a e b ( ( a s d ( s a i d e m f i e f f e f d r e a e r t n e a t n c t e r n e p t g r e i o r o v ) n i ) d e r ) What to Back Up Always: ...

A practical guide to backups and disaster recovery — automated backup strategies, testing your restores, and building systems that survive the worst.