Search suddenly not utilizing more than one core


(Do Hyung Kim) #1

Dear all,

I'm currently operating single node Elasticsearch setup on Ubuntu 16.04. It's Elasticsearch 5.4.0 from the very release date of it.

I had no problem at all with this setting till a couple of weeks ago. But from yesterday, I see searches are not leveraging all available cores (in my case, 4 cores).

The weirdest thing in this incident is I've not changed any configuration on the machine and my client code and machine. No ES configuration change and even no Ubuntu update (but after seeing the problem, I did update Ubuntu packages in the hope of updates may solve the problem).

Indexing levarages all cores. But searches don't.

I tried to set the log level to DEBUG and checked search thread pool is of size 7.

Is there anything I can do to find out what's going wrong?

The ES machine has:

  • Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz (4 core, no ht)
  • 8GiB RAM
  • Running Ubuntu 16.04 LTS

I assigned 4GiB RAM to JVM heap.
Since I'm using it for testing purpose only, it only runs benchmarks in controlled ways.

Thanks in advance.

Best regards,
DoHyung


(Do Hyung Kim) #3

It was due to a change in local network setting. I was not notified and didn't suspect network settings since indexing seemed ok.


(system) #4

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