literary just from your description, size of 50000 looks a lot. but it also depend on how many nodes you have in cluster, what are the existing work load like index/query.
How long is "long" in your case? Do you know if the machines are I/O or CPU bound while running this query? Could you try to capture the output of the hot threads API while the query is running, it could give an idea of where time is spent.
when the node is busy due to segment merging or gc on the heap, running this command it stuck in the command line forever, but in your case, it is worth to try.
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.