Repository navigation
ThreadPoolExecutor#shutdown? inconsistency in JRuby and C Ruby #1041
Description
Activity
Another discovery here is that in J Ruby, once
ThreadPoolExecutor#killis called, it's necessary to do await_for_terminationto achieve a fully shutdown and terminated executor.This is now the shutdown wrapper I'm using:
# @param timeout [Numeric, nil] Seconds to wait for active threads. # * +nil+, the scheduler will trigger a shutdown but not wait for it to complete. # * +-1+, the scheduler will wait until the shutdown is complete. # * +0+, the scheduler will immediately shutdown and stop any threads. # * A positive number will wait that many seconds before stopping any remaining active threads. def shutdown(timeout: -1) return if @executor.nil? || (@executor.shutdown? && !@executor.shuttingdown?) @executor.shutdown if @executor.running? if @executor.shuttingdown? && timeout # rubocop:disable Style/GuardClause executor_wait = timeout.negative? ? nil : timeout unless @executor.wait_for_termination(executor_wait) @executor.kill @executor.wait_for_termination end end end
This is similar to the usage example given in Java:
void shutdownAndAwaitTermination(ExecutorService pool) { pool.shutdown(); // Disable new tasks from being submitted try { // Wait a while for existing tasks to terminate if (!pool.awaitTermination(60, TimeUnit.SECONDS)) { pool.shutdownNow(); // Cancel currently executing tasks // Wait a while for tasks to respond to being cancelled if (!pool.awaitTermination(60, TimeUnit.SECONDS)) System.err.println("Pool did not terminate"); } } catch (InterruptedException ie) { // (Re-)Cancel if current thread also interrupted pool.shutdownNow(); // Preserve interrupt status Thread.currentThread().interrupt(); } }
Thanks! This line does look like the culprit:
I missed that last night because I got distracted trying to discover the need for the feature-check in
#shuttingdown?:That was introduced 9+ years ago (I couldn't find the exact place), but I'm not sure if
isTerminating/isTerminatedonly exist for certain implementations or versions of Java. Any idea how to check that?Aha, it looks like just
isTerminatingmaybe didn't exist in Java 6, butisTerminateddid. ok, I think I have enough for a PR 👍
In the scenario in which a ThreadPoolExecutor has received
#shutdownand there are active threads/tasks, there is an inconsistency between what JRuby and C Ruby return for#shutdown?:#shutdown? => truewhile there are still active threads/tasks, at the same time as#shuttingdown? => true. Java's executor service narrowly defines#isShutdownas "no new tasks will be accepted" which is distinct from#isTerminatedwhich is "all tasks have completed following shut down".#shutdown? => falsewhile there are still active threads/tasks, at the same time as#shuttingdown? => true.It's fairly easy to work around this inconsistency using
executor.shutdown? && !executor.shuttingdown?to achieve the same result in both rubies for determining when there are no more threads/tasks on a shutdown executor. It would be nice if the behavior was aligned.