Fragmentation on MongoDB databases

 

Fragmentation on MongoDB databases

When we removing a large amount of data from the collection and no plan to replace it, we generate a fragmentation, to deal with it, we have a COMPACT command.

The compact command rewrites and defragments all data and indexes in a collection, depending on what storage engine we have.

Differences between Storage – Engines (included in community edition)

COMMAND MMAPV1 WIREDTIGER
COMPACT On MMAPv1, compact defragments the collection’s data files and recreates its indexes. Unused disk space is not released to the system but instead retained for future data. On WiredTiger, compact attempts to reduce the required storage space for data and indexes in a collection, releasing unneeded disk space to the operating system.
COMPACT Compact requires up to 2 gigabytes of additional disk space to run on MMAPv1 databases. Compact does not require any additional disk space to run on WiredTiger databases
COMPACT  If you wish to reclaim disk space from an MMAPv1 database, you should perform an initial sync.

Considerations:

  • Always have an up-to-date backup before performing server maintenance such as the compact operation.
  • Compact only blocks operations for the database it is currently operating on. Only use compact during scheduled maintenance periods.
  • Compact command do not replicate to secondaries in a ReplicaSet
    • Compact each member separately
    • Ideally run compact on a secondary, this command forces to enter en recovering state
  • Shared Clusters
    • Compact only applies to a mongod instance, In a shared environment, run compact on each shard separately as a maintenance operation.
  • Index Building
    • mongod rebuilds all indexes in parallel following the compact operation
  • I recommend applying the compact operation if the result of the next formula is greater than 15%
    • 100% – (Data + Indexes) / storage > 15%
  • On databases larger than a few gigs can take a while to complete

Compacting and Compressing are two different things, please don´t confuse. These two concepts are totally different in MongoDB.

Daniel Cruz
Author: Daniel Cruz

Leave a Reply

Your email address will not be published. Required fields are marked *