[AMQ-496] Failover transport can cause CPU spin

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Cannot Reproduce
• Fix Version/s: 4.0 M4
• Component/s: Transport
• Assignee: Jonas B. Lim
• Reporter: Hiram R. Chirino
• Created: 18/Jan/06 23:18
• Updated: 16/Feb/06 11:30
• Resolved: 21/Jan/06 03:07

Description

4.0 - reliable:tcp transport (SNAPSHOT 01-12-06)
-----------------------------

• If you shutdown the ActiveMQ server, Consumers and Producers spin 100% cpu
  If you shutdown the ActiveMQ server, both my consumers and the producers will crank 100%. When the server is back, immediately everything will become okay again. I did a threadump of the JVM, and it wasn't blazingly obvious what was spinning all the cpu time. This is the closest I found:
  3XMTHREADINFO "ActiveMQ Default Thread Pool Thread" (TID:0x00002AAABF25EA00, sys_thread_t:0x00000000401101A8, state:CW, native ID:0x0000000040110208) prio=5
  4XESTACKTRACE at java/lang/Thread.sleep(Native Method)
  4XESTACKTRACE at java/lang/Thread.sleep(Thread.java:943)
  4XESTACKTRACE at org/apache/activemq/transport/failover/FailoverTransport$2.iterate(FailoverTransport.java:154)
  4XESTACKTRACE at org/apache/activemq/thread/SimpleTaskRunner.runTask(SimpleTaskRunner.java:110)
  4XESTACKTRACE at org/apache/activemq/thread/SimpleTaskRunner.access$100(SimpleTaskRunner.java:25)
  4XESTACKTRACE at org/apache/activemq/thread/SimpleTaskRunner$1.run(SimpleTaskRunner.java:42)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/Executors$RunnableAdapter.call(Executors.java:431)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/FutureTask.run(FutureTask.java:166)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:91)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:198)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:643)
  4XESTACKTRACE at edu/emory/mathcs/backport/java/util/concurrent/ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:668)
  4XESTACKTRACE at java/lang/Thread.run(Thread.java:816)

However that's clearly sleeping. But, either way, something on my consumers and producers is spinning 100% cpu each. It makes moving my mouse slow because of the thread priorities . Obviously a big big problem if you have many consumers/producers on a single machine. Probably a simple fix.