Backup & restore guides

Everything we know about not losing Postgres data — commands included, vendor pitches clearly marked.

• How to back up a Supabase database (3 ways)

  Manual pg_dump, Supabase's built-in backups, and automated verified backups — with the exact commands and the tradeoffs of each.

• Supabase free tier: you have no backups (fix it in 10 minutes)

  The Supabase free plan takes no automated backups. Here's the fastest path to a real, restorable backup — free options included.

• Fixing pg_dump/pg_restore version mismatch errors

  "server version mismatch" and "unsupported version in file header" explained — why they happen and the version-matching rules that prevent them.

• Restoring a Neon database: PITR, branches, and dumps

  Neon gives you three restore paths — point-in-time restore, branching from history, and plain pg_restore. When to use which.

• pg_dump formats: custom vs plain vs directory (pick custom)

  The four pg_dump output formats compared — compression, parallel restore, selective restore — and why custom format is the right default.

• How to actually test a Postgres backup restores

  A restore test is more than pg_restore exiting 0. The checklist: exit code, schema, row counts, index integrity, and query sanity.

• Encrypting Postgres backups properly (envelope encryption)

  Bucket-level SSE is not enough for database dumps. Client-side AES-256-GCM with envelope encryption, explained with commands.

• Designing a backup retention policy (GFS explained)

  How long to keep Postgres backups: the grandfather-father-son scheme, compliance drivers, and the one backup you must never delete.

• pg_restore errors, decoded (and which ones to ignore)

  role does not exist, must be owner of extension, already exists — the pg_restore errors that matter and the ones that are noise.

• Restoring a Supabase backup into a new Supabase project

  Migrating or recovering into a fresh Supabase project with pg_restore — auth schema, storage pointers, and the gotchas.

• Monitoring backups with a dead-man switch

  Backup jobs fail silently — cron won't tell you it stopped. Heartbeat monitoring (Healthchecks.io pattern) fixes the failure mode that matters most.

• RDS automated backups: what they cover and what they don't

  RDS snapshots and PITR are excellent — inside AWS. Cross-account risk, logical portability, and restore proof are the gaps.

• Using amcheck to detect index corruption

  Postgres indexes corrupt silently — queries return wrong answers, not errors. amcheck finds it. How to run it and what the errors mean.

• PITR vs logical backups: you need both

  WAL-based point-in-time recovery and pg_dump archives fail differently. The matrix of incidents shows why neither alone is enough.

• Backing up Railway Postgres (beyond volume snapshots)

  Railway snapshots restore disks, not data. Setting up pg_dump backups for Railway with the TCP proxy, plus a read-only role.

• Streaming pg_dump straight to S3/R2 (no local disk)

  Pipe pg_dump through encryption into aws s3 cp - for zero-disk backups: commands, multipart sizing, and failure handling.

• Creating a proper read-only Postgres role for backups

  The least-privilege role for pg_dump: grants, default privileges, the PUBLIC TEMP gotcha, and how to prove the role can't write.

• How long will your Postgres restore take? (RTO math)

  Restore time scales with data + index size, not dump size. How to measure your real RTO instead of guessing it during an outage.

• Backup evidence for SOC 2: what auditors actually ask

  Mapping backup practice to SOC 2 availability criteria: the policy, the evidence, and the restore-test question that fails teams.

• Neon branches are not backups (and when they are enough)

  Branches share the parent's storage and account. What branches solve brilliantly, and the three risks they can't cover.

• pg_dump vs pg_dumpall: globals, roles, and what you're missing

  pg_dump skips roles, tablespaces, and cluster settings. What pg_dumpall --globals-only covers and the combined backup recipe.

• Backing up Postgres running in Docker (without the footguns)

  docker exec pg_dump done right: version pinning, streaming out of the container, volume snapshots vs logical dumps, compose examples.

• Detecting schema drift with pg_dump hashes

  Hash pg_dump --schema-only output to detect unauthorized schema changes and verify restores byte-for-byte. Normalization tricks included.

• Backing up large Postgres databases (100GB+) without pain

  Directory-format parallel dumps, compression tradeoffs, replica-based dumping, and when logical backups stop scaling.

• The Postgres backup security checklist

  12 checks: credential scope, encryption at rest and in flight, key custody, access logging, deletion protection, and the plaintext audit.