mongodump

creates binary exports of MongoDB databases

TLDR

Dump entire database

$ mongodump --db [database]

Dump specific collection

$ mongodump --db [database] --collection [collection]

Dump with authentication

$ mongodump --uri "mongodb://[user]:[pass]@[host]/[db]"

Dump to specific directory

$ mongodump --out [/path/to/backup]

Dump as archive file

$ mongodump --archive=[backup.archive]

Dump compressed

$ mongodump --gzip --out [backup/]

Dump with query filter

$ mongodump --db [db] --collection [coll] --query '{"status": "active"}'

Dump from remote host

$ mongodump --host [hostname] --port [27017]

SYNOPSIS

mongodump [--uri uri] [--db database] [--collection coll] [--out dir] [--archive file] [options]

mongodump creates binary exports of MongoDB databases. It reads data directly from MongoDB and writes BSON files, preserving document structure and types.
Output format is a directory structure with one folder per database containing BSON and metadata JSON files for each collection. The archive format combines everything into a single file.
The --oplog option captures operations during the dump, enabling point-in-time restore. This is essential for consistent backups of active databases.
Query filtering exports only matching documents. This enables partial backups, test data extraction, or archiving specific records.
Compression (--gzip) significantly reduces backup size. Combined with archive format, it produces a single compressed file suitable for storage or transfer.
Parallelism options speed up large backups. Multiple collections can dump simultaneously, and documents within collections can be processed in parallel.

PARAMETERS

--uri URI

MongoDB connection URI.

--host HOST

Server hostname.

--port PORT

Server port.

--db, -d DATABASE

Database to dump.

--collection, -c COLL

Collection to dump.

--out, -o DIR

Output directory.

--archive[=FILE]

Output as archive file.

--gzip

Compress output.

--query, -q JSON

Filter documents.

--queryFile FILE

Query filter from file.

--username, -u USER

Authentication username.

--password, -p PASS

Authentication password.

--authenticationDatabase DB

Authentication database.

--oplog

Include oplog for point-in-time backup.

--dumpDbUsersAndRoles

Include users and roles.

--numParallelCollections N

Parallel collection dumps.

-j N

Parallel document dumps.

CAVEATS

Not recommended for large production databases - use filesystem snapshots instead. Backup is not atomic without --oplog. Indexes are saved as metadata, rebuilt on restore. Requires read access to all databases being dumped.

HISTORY

mongodump has been part of MongoDB since early versions, developed by MongoDB, Inc. (formerly 10gen). It provides the standard method for exporting MongoDB data, complemented by mongorestore for imports. It's part of the MongoDB Database Tools package.