I recently re-indexed my 14.7 million documents to a new index. Both source and destination index have 0 replicas. At the end of my re-indexing process, my destination index seems to occupy way more disk space than my source index? What could be the reason? (* please see the attached picture for clarity). Documents inside the parenthesis looks way more than the source as well
FYI: I used automatic and manual slicing multiple times since the process failed numerous times in between due to cluster failure, shards relocation, shards unavailability. I have also used “version conflict syntaxes” like this:
All of these are supposed to update the documents that are already indexed and create that are not. So, what could be the reason behind the uneven document count between two indexes? And what can be done in order to sort this out?
PICTURE OF SOURCE AND DESTINATION INDEX: