-
Bug
-
Resolution: Obsolete
-
Critical
-
None
-
8.2.1.Final
-
None
As a result of ISPN-6005, we decoupled the incoming Hot Rod server invocations from sending events by adding a intermediate queue that maintains the events to send to clients. However, this separation can lead to a Netty worker thread starvation issue since we add the events to the queue in Netty's IO thread, so if the queue is full, incoming requests are stuck.
[ISPN-6478] Netty worker thread starvation with events
Karsten Klein (Inactive)
added a comment - We seem to face the issues with higher load on our Infinispan Server based system.
Here the top most part of a stack dump from an .hprof. The thread never comes back and causes severe effects on our system (as entity locks are still in place):
java.lang.Thread @ 0x6c5ee2d70
- at sun.misc.Unsafe.park(ZJ)V (Native Method)
- at java.util.concurrent.locks.LockSupport.park(Ljava/lang/Object;)V (LockSupport.java:175)
- at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await()V (AbstractQueuedSynchronizer.java:2039)
- at java.util.concurrent.LinkedBlockingQueue.put(Ljava/lang/Object;)V (LinkedBlockingQueue.java:350)
- at org.infinispan.server.hotrod.ClientListenerRegistry$BaseClientEventSender.sendEvent(...)V (ClientListenerRegistry.scala:296)
- at org.infinispan.server.hotrod.ClientListenerRegistry$BaseClientEventSender.onCacheEvent(...)V (ClientListenerRegistry.scala:262)
- at sun.reflect.GeneratedMethodAccessor247.invoke(Ljava/lang/Object;[Ljava/lang/Object;)Ljava/lang/Object; (Unknown Source)
We are using 8.2.6 at the moment. Using the latest 8.2.x version is doable. Switching to 9.x at this point in time is not regarded an option.
How and when do you plan to approach this issue?
Karsten Klein (Inactive)
added a comment - We seem to face the issues with higher load on our Infinispan Server based system.
Here the top most part of a stack dump from an .hprof. The thread never comes back and causes severe effects on our system (as entity locks are still in place):
java.lang.Thread @ 0x6c5ee2d70
at sun.misc.Unsafe.park(ZJ)V (Native Method)
at java.util.concurrent.locks.LockSupport.park(Ljava/lang/Object;)V (LockSupport.java:175)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await()V (AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.put(Ljava/lang/Object;)V (LinkedBlockingQueue.java:350)
at org.infinispan.server.hotrod.ClientListenerRegistry$BaseClientEventSender.sendEvent(...)V (ClientListenerRegistry.scala:296)
at org.infinispan.server.hotrod.ClientListenerRegistry$BaseClientEventSender.onCacheEvent(...)V (ClientListenerRegistry.scala:262)
at sun.reflect.GeneratedMethodAccessor247.invoke(Ljava/lang/Object;[Ljava/lang/Object;)Ljava/lang/Object; (Unknown Source)
We are using 8.2.6 at the moment. Using the latest 8.2.x version is doable. Switching to 9.x at this point in time is not regarded an option.
How and when do you plan to approach this issue? 
Infinispan issue tracking has been migrated to GitHub issues: https://github.com/infinispan/infinispan/issues
If you still want this issue to be worked on, create a new issue on GitHub and link this issue.