Why one of nodes rejected more requests?

Elasticsearch cluster has 2 nodes. But when I start stress test using esrally, I found that one of them has more rejections than the other
like this:

node_name name active queue rejected completed
elasticsearch-logging-0 write 0 0 108266 70186
elasticsearch-logging-1 write 0 0 5662 70109

  1. Each of them has 2 cores, but their bulk thread pool processor is 2.
  2. My rally clients number is 200 and the bulk_size is 100.
  3. Each doc is about 900 bytes.
  4. The error rate from esrally is 95%!!!!!!!!

PLZ! Can anyone help me!!! ORZ!!!

I think may be I should add a coordinating only node to load banlance?

But how to change elasticsearch config file when I use kubernetes to manage es cluster?

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