-
Notifications
You must be signed in to change notification settings - Fork 100
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
enable to preload common pool #199
enable to preload common pool #199
Conversation
private void preloadCommonPool() | ||
{ | ||
try | ||
{ | ||
// ensure common pool exists in the jvm | ||
final Class<?> fjp = Thread.currentThread().getContextClassLoader() | ||
.loadClass("java.util.concurrent.ForkJoinPool"); | ||
final ExecutorService es = ExecutorService.class.cast( | ||
fjp | ||
.getMethod( "commonPool" ) | ||
.invoke( null ) ); | ||
final int max = preloadCommonPool > 0 | ||
? preloadCommonPool : | ||
Integer.class.cast( fjp.getMethod( "getCommonPoolParallelism" ).invoke( null ) ); | ||
final CountDownLatch preLoad = new CountDownLatch( 1 ); | ||
for ( int i = 0; | ||
i < max; | ||
i++ ) | ||
{ | ||
es.submit(new Runnable() { | ||
@Override | ||
public void run() { | ||
try | ||
{ | ||
preLoad.await(); | ||
} | ||
catch ( InterruptedException e ) | ||
{ | ||
Thread.currentThread().interrupt(); | ||
} | ||
} | ||
}); | ||
} | ||
preLoad.countDown(); | ||
} | ||
catch (final Exception e) | ||
{ | ||
getLog().debug(e.getMessage() + ", skipping commonpool earger init"); | ||
} | ||
} | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
After our chat, I think I understood what is going on here. I am not sure, however, why reflection would be needed.
Anyway. Most of us might not be that familiar with what is going on here. A few words on the 'why' would help a lot:
why using reflection, why initializing it (maybe in conjunction with the ThreadGroup).
From my understanding, no reflection would be needed because the created JFP would not be part of the thread group.
Code comments on the "why" would be great for our maintainers! :)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Moved to java 8 since plugin does not use java 7 anymore which drops the reflection need.
Code by itself does not use thread groups but they are inherited from the calling thread so if commonpool is created lazily in the context of the main (isolatedclassloader) then it would use this threadgroup in some JVM which would let the JVM hang.
This is what https://github.com/mojohaus/exec-maven-plugin/pull/199/files#diff-dd2fe98649dca0e18d01909e1fc59b14c562cf349cb49a2cf19957147dca5551R76 means, happy to get a better wording if you find one, also not sure of the default we should use but at least we have a clean way to solve that issue now.
Superseeded by #279 |
related to #198
Long story short, since exec:java tries to stop new threads, commonPool ones are identified as such if not already created, issue is that if the exec:java uses commonPool, potentially not in the code (thinking to HttpClient which always uses commonPool for part of its tasks), then exec:java will not stop properly until you skip the shutdown cleanup/daemon flags which is not desired. So idea is to enable to preload the common pool and skip these JVM threads.