All snapshots are full. The incremental part refers to only new segments bring copied every time a new snapshot is created. This very old blog post describes it quite well and is still relevant as far as I know.
I also wrote a blog on how this works with different levels of detail, including deletes.
I do think it's confusing to call snapshots 'full', as to me they are incremental, but restores are always full restores. It's different than how DBMS do incremental, but to me any backup that only 'writes' changed data is incremental.
Maybe that's just from being an old DBA, but I bet this confuses a lot of folks and kinda wish the nomenclature were different (incremental snaps, full restores).
Thanks and happily the latest ES docs use incremental terminology:
Snapshots are incremental: each snapshot only stores data that is not part of an earlier snapshot. This enables you to take frequent snapshots with minimal overhead.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.