jca workmanager max-threads does not seem to work
jim_b_o Dec 1, 2015 11:20 AMWhen configuring the jca workmanager long-running-threads I'm setting "core-threads" => 60 and "max-threads" => 120 but the pool never grows above 60. Per the output below we can see that there are 60 threads in use and 2 more jobs queued that never get started.
/subsystem=jca/workmanager=default/long-running-threads=default:read-resource(recursive=true,include-runtime=true)
{
"outcome" => "success",
"result" => {
"allow-core-timeout" => false,
"core-threads" => 60,
"current-thread-count" => 60,
"handoff-executor" => undefined,
"keepalive-time" => {
"time" => 10L,
"unit" => "SECONDS"
},
"largest-thread-count" => 60,
"max-threads" => 120,
"name" => "default",
"queue-length" => 50,
"queue-size" => 2,
"rejected-count" => 0,
"thread-factory" => undefined
}
}
/subsystem=jca/workmanager=default/statistics=local:read-resource(recursive=true,include-runtime=true)
{
"outcome" => "success",
"result" => {
"dowork-accepted" => 0,
"dowork-rejected" => 0,
"schedulework-accepted" => 0,
"schedulework-rejected" => 0,
"startwork-accepted" => 62,
"startwork-rejected" => 0,
"statistics-enabled" => true,
"work-active" => 60,
"work-failed" => 0,
"work-successful" => 60,
"workmanager-statistics-enabled" => true
}
}
If I set "core-threads" => 120 then everything is OK as shown below.
/subsystem=jca/workmanager=default/long-running-threads=default:read-resource(recursive=true,include-runtime=true)
{
"outcome" => "success",
"result" => {
"allow-core-timeout" => false,
"core-threads" => 120,
"current-thread-count" => 87,
"handoff-executor" => undefined,
"keepalive-time" => {
"time" => 10L,
"unit" => "SECONDS"
},
"largest-thread-count" => 87,
"max-threads" => 120,
"name" => "default",
"queue-length" => 50,
"queue-size" => 0,
"rejected-count" => 0,
"thread-factory" => undefined
}
}
/subsystem=jca/workmanager=default/statistics=local:read-resource(recursive=true,include-runtime=true)
{
"outcome" => "success",
"result" => {
"dowork-accepted" => 0,
"dowork-rejected" => 0,
"schedulework-accepted" => 0,
"schedulework-rejected" => 0,
"startwork-accepted" => 87,
"startwork-rejected" => 0,
"statistics-enabled" => true,
"work-active" => 87,
"work-failed" => 0,
"work-successful" => 87,
"workmanager-statistics-enabled" => true
}
}
So it looks like max-threads is being ignored. Perhaps this is similar to https://issues.jboss.org/browse/WFLY-5332?
This is occurring with WildFly 10.0.0.CR4 and 10.0.0.CR5-SNAPSHOT from about a week ago.
So the workaround is easy but it looks like there may be a bug. Can anyone confirm having similar problems?