I am running Logstash as a service on a 16 core machine, with following conf
curl -XGET 'localhost:9600/_node/pipelines?pretty'
{
"host" : "0.0.0.0",
"version" : "6.1.2",
"http_address" : "127.0.0.1:9600",
"id" : "8311cdf3-165c-47b4-8373-37cd2007d858",
"name" : "0.0.0.0",
"pipelines" : {
"main" : {
"workers" : 16,
"batch_size" : 100,
"batch_delay" : 5,
"config_reload_automatic" : false,
"config_reload_interval" : 3000000000,
"dead_letter_queue_enabled" : false
}
}
However the machine has a "isolated cpu" setting : isolcpus=0-13
I have defined the CPUAffinity for logstash via systemctl as : CPUAffinity=0-13
My problem is that all the logstash pipeline worker threads are still running on core 0, but not being distributed among all the available cores. Please help.