Version: 2.6.0
Describe the issue:
What is the equivalent for threadpool bulk writes and reads in Opensearch?
I found these by running querying the default settings
“threadpool”:{“queue_size”:“200”,“size”:“4”}
Configuration:
This is what I have in older versions of ES, wanted to see if there is anything similar to that
"threadpool":{"bulk":{"queue_size":"2500"},"index":{"queue_size":"500"}
see thread_pool.search
:
"search": { "max_queue_size": "1000", "queue_size": "1000", "size": "8", "auto_queue_frame_size": "2000", "target_response_time": "1s", "min_queue_size": "1000" },
, and thread_pool.write
:
"write": { "queue_size": "10000", "size": "5" },
the default queue sizes are 1000 and 10000, which are same to ES 7.10.
1 Like
@gaobinlong
I checked my default settings, the only thing I was able to find for thread_pool where these
thread_pool":{“queue_size”:“1000”,“size”:“0”}
“thread_pool”:{“force_merge”:{“queue_size”:“-1”,“size”:“1”}
Does that mean I have to set this up? Do I set it up in my config file, or update it as persistent setting via API?
That’s wired, did you get the default settings by GET _cluster/settings?include_defaults=true
?
Correct. What version of OpenSearch do you have? 2.6.0
@stecino yes, it’s 2.6.0,
1 Like
I found these in my json
"search": {
"max_queue_size": "1000",
"queue_size": "1000",
"size": "7",
"auto_queue_frame_size": "2000",
"target_response_time": "1s",
"min_queue_size": "1000"
}
"write": {
"queue_size": "10000",
"size": "4"
}
Thanks
1 Like