Nodes with CPU high running Elasticsearch 5.2.2, suspect it to be segments merging issue as we found a huge number of deleted document, which might be running slow and causing CPU to hang on to 98-99%?

Elastic Stack Elasticsearch
1

We have (4core,32gb, and 600GB SSD x 9 machines) with index-A 15 Primary and 2Replicas and 2355 segments with the highest segment sizing 6.8gb, Index-B 15 Primary and 2Replicas and 1354 segments with the highest segment sizing 7.8gb and Index-C 15 Primary and 2Replicas and 1971 segments with the highest segment sizing 5.5gb. Have few queries with respect to resolving and finding the root cause of the issue. Any advice would be really grateful.

  1. What are the steps to run a force merge on a read-write index? Pros and Cons of merging segments.
  2. What would be an ideal number to be set for max_number_segment? Or any formula to calculate the same.
  3. Can I re-index a particular shard in Elasticsearch? Will it merge index automatically?
  4. What is better to be carried out on an index, reindexing or force merging(optimize)?
  5. Will the replica be usable during my reindex or merge segment operation?
  6. Is it safe to cancel running merge segment task?
2

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.