Rollbacks During Replica Set Failover¶
A rollback reverts write operations on a former primary when the member rejoins its replica set after a failover. A rollback is necessary only if the primary had accepted write operations that the secondaries had not successfully replicated before the primary stepped down. When the primary rejoins the set as a secondary, it reverts, or “rolls back,” its write operations to maintain database consistency with the other members.
MongoDB attempts to avoid rollbacks, which should be rare. When a rollback does occur, it is often the result of a network partition. Secondaries that can not keep up with the throughput of operations on the former primary, increase the size and impact of the rollback.
A rollback does not occur if the write operations replicate to another member of the replica set before the primary steps down and if that member remains available and accessible to a majority of the replica set.
Collect Rollback Data¶
When a rollback does occur, administrators must decide whether to apply or ignore the rollback data. MongoDB writes the rollback data to BSON files in the rollback/ folder under the database’s dbPath directory. The names of rollback files have the following form:
Administrators must apply rollback data manually after the member completes the rollback and returns to secondary status. Use bsondump to read the contents of the rollback files. Then use mongorestore to apply the changes to the new primary.
Avoid Replica Set Rollbacks¶
To prevent rollbacks, use replica acknowledged write concern to guarantee that the write operations propagate to the members of a replica set.
A mongod instance will not rollback more than 300 megabytes of data. If your system must rollback more than 300 megabytes, you must manually intervene to recover the data. If this is the case, the following line will appear in your mongod log:
[replica set sync] replSet syncThread: 13410 replSet too much data to roll back
In this situation, save the data directly or force the member to perform an initial sync. To force initial sync, sync from a “current” member of the set by deleting the content of the dbPath directory for the member that requires a larger rollback.