WiredTiger of MongoDb has redistribution performance issue as MMAPv1

Question

As MMAPv1 Document said

All records are adjacent to the disk, and when the document becomes larger than the selected record, MongoDB must select a new record. New distributions require that MongoDB move the document and update all indexes related to the document, which takes longer than in-place updates and results in memory fragmentation. Changed in version 3.0.0.
By default, MongoDB uses Power of 2 Sized Allocations so that each document in MongoDB is stored in a record that contains the document itself and additional space or addition. Filling allows the document to grow as a result of updates while minimizing the probability of redistribution.

But the WiredTiger Document says nothing about this. So I just want to know if it’s very good when the record size changes or has some performance problems, but is not mentioned in the document.

Answer 1

You do not need to worry about moving documents, filling out, etc. using WiredTiger. New entries are first written to files in unused regions, and then included with the rest of the data in the background later. During the upgrade, WiredTiger will actually write a new version of the documents, rather than overriding existing data, as mmapv1 does in many cases. (Check out the video from MongodDB's free online courses)