Hello! I'm looking into the performance of the delete index API and had a few questions:
How fast is deleting an Elasticsearch index? Any estimates on how long it would take to delete an index with 100s of GB of data? Or its performance on deleting an index with millions of objects?
What happens under the hood when the call to the delete index API is made? Are indexed objects deleted one-by-one?
Let me know if there's an online reference for this kind of information! I tried Googling around but didn't find much.
Thank you Mark! No troubleshooting, but I am currently using Elasticsearch as a search index, and am building out some tools to support its use. I have some concerns on deleting a very large Elasticsearch index.
It might be helpful to get a short blurb on how it operates. For example, even if the delete of an index is very fast / lightweight, if it deletes docs one by one before returning that may still be an issue for me as the count of indexed objects / total indexed size continues to grow.
It does not delete documents individually so is very fast and efficient. As the whole index and related files are deleted the cluster state need to be updated and propagated to all nodes. Having a large number of shards and aNd a large cluster state can slow this down but it should still be quite fast.
This is not something that is usually a performance problem so it would be interesting to know the rationale/reason behind this question.
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.