crdb-dump

A feature-rich CLI for exporting and importing CockroachDB schemas and data. Includes support for parallel chunked exports, manifest checksums, BYTES/UUID/ARRAY/VECTOR types, multi-schema (non-public) objects, permission introspection, secure resumable imports, S3-compatible storage (MinIO, Cohesity), region-aware filtering, and automatic retry logic.

Requires Python 3.10+.

📖 Documentation: https://viragtripathi.github.io/crdb-dump/

⚠️ Breaking changes in 0.4.0

• All object names are now three-part database.schema.table (filenames, manifests, resume-log keys, and --tables input). Objects in non-public schemas are now exported and restored correctly.
• --tables two-part input means schema.table (database taken from --db), not the old database.table. Use db.schema.table to be explicit, or a bare table for the public schema.
• Data chunk files are now db.schema.table_NNN.csv|sql; manifests are db.schema.table.manifest.json. Pre-0.4.0 dumps are not compatible.

See CHANGELOG.md for the full list.

🚀 Features

• ✅ Schema export: tables, views, sequences, enums (objects in any schema, not just public)
• ✅ Full-database dumps use native SHOW CREATE ALL TABLES/ALL TYPES (dependency-ordered, FK constraints validated post-load)
• ✅ Data export: CSV or SQL with chunking, gzip, and ordering
• ✅ Types: handles BYTES, UUIDs, STRING[], TIMESTAMP, enums, VECTOR
• ✅ Schema output formats: sql, json, yaml
• ✅ Resumable COPY-based imports with chunk-level tracking
• ✅ Permission exports: roles, grants, role memberships
• ✅ Parallel loading (--parallel-load) and manifest verification
• ✅ Dry-run for schema or chunk loading
• ✅ TLS and insecure auth supported
• ✅ Schema diff support (--diff)
• ✅ Full logging via logs/crdb_dump.log
• ✅ Automatic retry logic with exponential backoff for transient failures
• ✅ Fault-tolerant, resumable imports with --resume-log or --resume-log-dir
• ✅ Region-aware export/import via --region
• ✅ S3-compatible support (--use-s3) with MinIO, Cohesity, or AWS
• ✅ CSV header validation (--validate-csv)
• ✅ Python-based S3 bucket creation (via boto3) for MinIO

📦 Installation

pip install crdb-dump

🧪 Local Testing

./test-local.sh

This script will:

• Start a multi-region demo CockroachDB cluster
• Create test schema + data
• Export schema and chunked data (CSV)
• Verify chunk checksums
• Dry-run and real import with retry/resume
• Upload chunks to MinIO (S3-compatible)
• Download and verify import from S3
• Use Python (boto3) to create S3 buckets

🔧 CLI Overview

crdb-dump --help
crdb-dump export --help
crdb-dump load --help

Example usage:

crdb-dump export --db=mydb --data --per-table
crdb-dump load --db=mydb --schema=... --data-dir=... --resume-log=resume.json

🔐 Connection

export CRDB_URL="cockroachdb://root@localhost:26257/defaultdb?sslmode=disable"
# or
export CRDB_URL="postgresql://root@localhost:26257/defaultdb?sslmode=disable"

Alternatively:

--db mydb --host localhost --certs-dir ~/certs

Use --print-connection to verify resolved URL.

🏗 Export Options

crdb-dump export \
  --db=mydb \
  --per-table \
  --data \
  --data-format=csv \
  --chunk-size=1000 \
  --data-order=id \
  --data-compress \
  --data-parallel \
  --verify \
  --include-permissions \
  --archive

Schema Output

Data Export

⛓ Import Options

crdb-dump load \
  --db=mydb \
  --schema=crdb_dump_output/mydb/mydb_schema.sql \
  --data-dir=crdb_dump_output/mydb \
  --resume-log=resume.json \
  --validate-csv \
  --parallel-load \
  --print-connection

🔄 Fault Tolerance & Resume Support

• ✅ Retries failed operations with exponential backoff

• ✅ Resumable imports:

  • --resume-log (single file)
  • --resume-log-dir (per-table)
  • --resume-strict (abort on failure)

Writes resume state after each successful chunk. Restarts are safe and idempotent.

☁️ S3 / MinIO / Cohesity Example

crdb-dump export \
  --db=mydb \
  --per-table \
  --data \
  --chunk-size=1000 \
  --data-format=csv \
  --use-s3 \
  --s3-bucket=crdb-test-bucket \
  --s3-endpoint=http://localhost:9000 \
  --s3-access-key=minioadmin \
  --s3-secret-key=minioadmin \
  --s3-prefix=test1/ \
  --out-dir=crdb_dump_output

crdb-dump load \
  --db=mydb \
  --data-dir=crdb_dump_output/mydb \
  --resume-log-dir=resume/ \
  --parallel-load \
  --validate-csv \
  --use-s3 \
  --s3-bucket=crdb-test-bucket \
  --s3-endpoint=http://localhost:9000 \
  --s3-access-key=minioadmin \
  --s3-secret-key=minioadmin \
  --s3-prefix=test1/

🔍 Schema Diff Example

crdb-dump export --db=mydb --diff=old_schema.sql

Output:

crdb_dump_output/mydb/mydb_schema.diff

🧪 Testing

Requires Python 3.10+.

python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"

# Unit tests (no database needed)
pytest -m "not integration"

# Integration tests (need a reachable CockroachDB)
export CRDB_URL="cockroachdb://root@localhost:26257/defaultdb?sslmode=disable"
pytest -m integration

# Full end-to-end (needs cockroach + Docker/MinIO)
./test-local.sh

🚀 Releasing (maintainers)

Releases publish to PyPI via the Release GitHub Action (.github/workflows/release.yml) using PyPI Trusted Publishing (OIDC) — no API tokens stored in the repo.

One-time setup on PyPI: add a Trusted Publisher for the crdb-dump project → owner viragtripathi, repository crdb-dump, workflow release.yml.

To cut a release:

1. Bump version in pyproject.toml and update CHANGELOG.md; merge to main.
2. Run the Release workflow (Actions → Release → Run workflow) and enter the same version (e.g. 0.4.0).

The workflow verifies the input matches the packaged version, runs the full test suite against a CockroachDB container, builds the sdist/wheel, publishes to PyPI, and creates a v<version> GitHub Release with auto-generated notes.

❤️ Contributing

Pull requests welcome! Star ⭐ the repo, file issues, or request features at:

👉 https://github.com/viragtripathi/crdb-dump/issues