Elasticsearch document removal not removing all that was added

Question

I have been doing some analysis on the Elasticsearch storage and indexing of Documents and its impact in the physical memory.

Scenario:

Step1: The "data" folder in ym ES node is of size 48KB.

Step2: A file of size 100KB is added to the my ES. The "data" folder size increased to 410KB

Step3: Delete the document. The "data" folder size decreased to 345KB.

Question: Why does the size of the "data" folder is still 345KB. I assumed it should be 48KB because the document is deleted.

Answer 1

Most of the data is just going to be flagged for deletion until the deletions are flushed. That usually happens when enough deletions have occurred to perform that operation (which is generally expensive).

I believe the decrease you are seeing immediately when deleting that document is just the stored portion of it, but not the indexed terms, etc.