Ok. Is there a rough estimate of how long it should take? i.e. 100 segments each around 1Gb should take NN seconds to force merge down to 1 segment?
From this guide page I see that heavy merging is not throttled on IO and that it should be done on dedicated "merge" host. What does this mean for replicas? Does merging get done on the primary first and then copied to replicas dropping existing replicas? Or does the merge occur on any host that holds an active copy of the shard (replica or primary) meaning that I need to allocate both primaries AND replicas to "dedicated force merge" hosts?
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.