MongoDB Plugin: Scope
EnterpriseBacula Enterprise Only
This solution is only available for Bacula Enterprise. For subscription inquiries, please reach out to sales@baculasystems.com.
The Bacula Enterprise MongoDB Plugin is designed for Linux File Daemon hosts
that can reach the MongoDB deployment and can run the MongoDB Database Tools,
specifically mongodump and mongorestore. The plugin does not require
mongosh for backup or restore operations.
The plugin is available starting with Bacula Enterprise 18.2.5.
Supported MongoDB Server Versions
At the time of the plugin release, the following MongoDB server versions are supported: 6.0, 7.0, and 8.0. Future official versions are expected to be supported by the Bacula Enterprise MongoDB Plugin.
For more information about supported MongoDB server platforms and versions, contact Bacula Systems Support Team.
Restore to the same MongoDB major version, or to a newer major version, is supported, subject to MongoDB compatibility and feature compatibility version checks. Downgrade restores to an older MongoDB major version are rejected by the plugin before restore starts.
Supported Deployment Types
The plugin supports the following deployment types:
Standalone
mongodinstancesWritable replica-set primaries
Sharded clusters accessed through
mongos.
The plugin intentionally does not support direct backup of secondaries, direct
backup of shard members, or direct backup of config servers. For sharded
clusters, always connect to mongos, the MongoDB query router. This is not
mongosh, the interactive MongoDB shell.
Supported Security Modes
The plugin supports:
No-authentication MongoDB deployments
SCRAM username/password authentication
Passwords provided directly, through a password file, or through an environment variable
TLS server validation with a CA file
TLS client certificate/key files
MONGODB-X509authentication using a client certificate identity.
Supported Backup Levels and Recovery Semantics
Bacula Full, Incremental, and Differential job levels are accepted. Their behavior depends on the source topology:
Source topology |
Full |
Incremental |
Differential |
Recovery semantics |
|---|---|---|---|---|
Standalone |
Full logical artifact |
Full-style artifact |
Full-style artifact |
Restore from the selected logical artifact. Point-in-time replay is not available because there is no replica-set oplog recovery chain. |
Writable replica-set primary |
Full logical artifact with base oplog window |
Oplog-chain segment from the previous chain boundary |
Oplog-chain segment from the latest Full boundary |
Full plus Incremental or Differential artifacts can be restored as a validated oplog chain and replayed up to the selected recovery boundary. |
Sharded cluster through |
Coordinated logical full artifact with routing metadata |
Full-style artifact |
Full-style artifact |
Restore through guardrails. Sharded oplog point-in-time recovery is not supported. |
Direct replica-set secondary, shard member, or config server |
Not supported |
Not supported |
Not supported |
Connect to a writable replica-set primary for replica sets, or to
|
The entries in the matrix mean:
Full logical artifact: a logical MongoDB backup produced with MongoDB Database Tools and described by the plugin manifest. It contains the selected databases, collections, indexes, and eligible security metadata.
Full logical artifact with base oplog window: a Full replica-set backup that also records the oplog boundary needed to start a later recovery chain.
Coordinated logical full artifact with routing metadata: a sharded-cluster logical backup taken through
mongos. The plugin captures sharded planning metadata and enforces balancer guardrails, but it does not replace config server disaster recovery.Full-style artifact: an Incremental or Differential Bacula job that is accepted for schedule continuity, but is executed as a complete logical artifact for that job. It is safe to restore, but it is not an oplog delta and does not provide point-in-time replay.
Oplog-chain segment: a replica-set Incremental or Differential payload containing the validated
local.oplog.rsrange since the previous recovery boundary. Restore replays these segments after the base Full artifact.Recovery boundary: the oplog timestamp where replay stops. The plugin can auto-select the chain tip or use an explicit operator-provided replay point when it falls inside the protected oplog range.
Use-Case Callouts
Application teams needing reliable logical MongoDB recovery without scripting.
Backup administrators who need MongoDB data in Bacula retention policies and storage pools.
Regulated environments requiring off-site, tape, or air-gapped copies of MongoDB data.
Replica-set deployments requiring incremental recovery chains and point-in-time replay.
Sharded deployments where cluster routing metadata must be captured and used during restore planning.
See also
Next articles:
Go back to: MongoDB Plugin.