[
https://issues.apache.org/jira/browse/POOL-386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17156799#comment-17156799
]
Phil Steitz commented on POOL-386:
----------------------------------
I am sorry. I was not clear. What I implemented was just usage tracking, with
the efficiency consideration you mentioned. I added tests to confirm that as
long as all schedules are followed by cancels (which will happen if pools are
always closed), the executor is properly shut down. Looking more carefully
this AM, I think my comment above about the current setup is likely incorrect.
When I said "the current setup may benefit" I meant it already had the benefit,
but I think now that is likely incorrect.
Your idea looks interesting as a way to actually handle the case of abandoned
pools.
> Closing a pool can cause Evictor in another pool to be cancelled
> ----------------------------------------------------------------
>
> Key: POOL-386
> URL: https://issues.apache.org/jira/browse/POOL-386
> Project: Commons Pool
> Issue Type: Task
> Affects Versions: 2.6.1, 2.6.2, 2.7.0, 2.8.0
> Reporter: Phil Steitz
> Priority: Major
> Fix For: 2.8.1
>
>
> The fix for POOL-337 introduced a race condition that can cause the shared
> EvictionTimer to be cancelled when a pool is closed but another pool still
> has an active Evictor. The EvictionTimer cancel method used to cancel an
> eviction task owned by a client pool uses this test to determine whether or
> not to shutdown its executor:
> {code:java}
> if (executor != null && executor.getQueue().isEmpty()){code}
> The executor may report an empty queue if it is executing a task. This will
> cause the executor to be shut down and scheduling of new tasks to stop.
> The unit test below illustrates the problem.
> {code:java}
>
> public void testEvictionTimerMultiplePools() throws InterruptedException {
> final AtomicIntegerFactory factory = new AtomicIntegerFactory();
> factory.setValidateLatency(50);
> final GenericObjectPool<AtomicInteger> evictingPool = new
> GenericObjectPool<>(factory);
> evictingPool.setTimeBetweenEvictionRunsMillis(100);
> evictingPool.setNumTestsPerEvictionRun(5);
> evictingPool.setTestWhileIdle(true);
> evictingPool.setMinEvictableIdleTimeMillis(50);
> for (int i = 0; i < 10; i++) {
> try {
> evictingPool.addObject();
> } catch (Exception e) {
> e.printStackTrace();
> }
> }
> for (int i = 0; i < 1000; i++) {
> GenericObjectPool<AtomicInteger> nonEvictingPool = new
> GenericObjectPool<>(factory);
> nonEvictingPool.close();
> }
> Thread.sleep(1000);
> Assert.assertEquals(0, evictingPool.getNumIdle());
> evictingPool.close();
> }
> {code}
> Proposed fix:
> Change the test guarding executor shutdown to
> {code:java}
> executor.getQueue().isEmpty() && executor.getActiveCount() == 0
> {code}
> This still exposes a low-probability race - a task completes after the
> isEmpty test and is requeued before the activecount test - but this is very
> unlikely.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
