Slow query in a singal server with big data for the first query

I'm using elasticsearch 5.6.2 in a signal server and give heap memory for elasticsearch is 16GB, the disk is SAS 10krpm。

The indices with 5 shards and 1 replica by default.

1 day data for one indices is about 20GB (the docs number is about 60 million to 300 million) or more bigger in elasticsearch, when i do simple query from the GUI, it will take more than 10s (sometimes 40s) to get the data for the first query, but the second query will be fast, just take 3s。 I also monitor the disk IO, and see that the disk io util is 100% during the first query, but the heap memory usage by elasticsearch is slow(about 5GB(get from bigdesk), more less than given 16GB).

So, my question is how to improve the query performance in a singal server with big indices?

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