Multiple core cpu

I have problems with multiple cpu cores use.
My server have 12cores and available everytime, but on the peak time it be stuck because Elasticserach cannot use more than one core.
I check process by top command and see only one core have load 150-200%
Cannot change config to use more than one core :frowning:

What is your use case? Is it indexing or query heavy? How many shards and indices do you have? Is there anything else limiting performance, e.g. disk I/O?

I have ubuntu server with 12 cores cpu . 70GB RAM. SSD Disk
On this server i running

  • nginx server
  • ELK
  • uwsgi django application
    I config for nginx and uwsgi to use maximum of cores. On the peak time, almost cpu have load < 60% but only one core run elasticsearch go upto 200%.
    Stast info
    "_shards": {
    "total": 32,
    "successful": 16,
    "failed": 0
    }

What indexing and/or query rate do you need to support? Are you throwing enough load at the node to make it require the use of more than 1-2 cores?

I will explain our purpose.
I have 1 main index for product information, and 2 logs indexes. When user request to view product list, i will query for search on elasticsearch and return information, before return info i log request information to elasticsearch.
On peak time, for example when we send push notification to user, user access to out system, as i imagine, each request will be process by one uwsgi server and on this time it will make 2 request to elastisearch system to search and write logs.
I have multiple uwsgi (24) process running to handle request, so as current situation only one elasticsearch process is running so it make request slow down because of queuing for elasticsearch. so if we can make it running on more than 1 core it will reduces response time

What makes you think that Elasticsearch can only use one core?

1 Like

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