[
https://issues.apache.org/jira/browse/IGNITE-22959?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17956122#comment-17956122
]
Kirill Tkalenko commented on IGNITE-22959:
------------------------------------------
Looks good.
> Fix incorrect ThreadPoolExecutor usages
> ----------------------------------------
>
> Key: IGNITE-22959
> URL: https://issues.apache.org/jira/browse/IGNITE-22959
> Project: Ignite
> Issue Type: Improvement
> Reporter: Philipp Shergalis
> Assignee: Philipp Shergalis
> Priority: Major
> Labels: ignite-3
> Time Spent: 1h
> Remaining Estimate: 0h
>
> ThreadPoolExecutors with unbounded queues and without allowCoreThreadTimeout
> actually work as fixed thread pools, ignoring maximumPoolSize and keepAlive
> properties. All usages of ThreadPoolExecutor constructor should be examined
> and fixed:
> 1) If this behaviour is expected, replace constructor with
> newFixedPoolExecutor() to avoid confusion
> 2) If pool is expected to resize only from corePoolSize to maximumPoolSize,
> SynchronousQueue can be used - but it will reject tasks if maximum amount of
> threads is already created. We can use Integer.MAXIMUM as maximumPoolSize or
> introduce a custom queue to avoid this
> 3) If pool is expected to resize from 0 to maximumPoolSize,
> allowCoreThreadTimeout should be used
>
> Example of incorrect usage:
> {code:java}
> tableIoExecutor = new ThreadPoolExecutor( Math.min(cpus * 3, 25),
> Integer.MAX_VALUE, 100, MILLISECONDS, new LinkedBlockingQueue<>(),
> IgniteThreadFactory.create(nodeName, "tableManager-io", LOG, STORAGE_READ,
> STORAGE_WRITE));{code}
> From ThreadPoolExecutor documentation:
> (for example a LinkedBlockingQueue without a predefined capacity will cause
> new tasks to wait in the queue when all corePoolSize threads are busy. Thus,
> no more than corePoolSize threads will ever be created. (And the value of the
> maximumPoolSize therefore doesn't have any effect.)
> By default, the keep-alive policy applies only when there are more than
> corePoolSize threads, but method #allowCoreThreadTimeOut(boolean) can be used
> to apply this time-out policy to core threads as well, so long as the
> keepAliveTime value is non-zero.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)