[HBASE-14274] Deadlock in region metrics on shutdown: MetricsRegionSourceImpl vs MetricsRegionAggregateSourceImpl

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.2.0, 1.3.0, 2.0.0
• Component/s: test
• Hadoop Flags: Reviewed
• Assignee: Elliott Neil Clark
• Reporter: Michael Stack
• Created: 20/Aug/15 19:17
• Updated: 01/Jul/22 20:59
• Resolved: 21/Aug/15 03:48

Description

Looking into parent issue, got a hang locally of TestDistributedLogReplay.

We have region closes here:

"RS_CLOSE_META-localhost:59610-0" prio=5 tid=0x00007ff65c03f800 nid=0x54347 waiting on condition [0x000000011f7ac000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x000000075636d8c0> (a java.util.concurrent.locks.ReentrantReadWriteLock$NonfairSync)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:834)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireQueued(AbstractQueuedSynchronizer.java:867)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquire(AbstractQueuedSynchronizer.java:1197)
	at java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.lock(ReentrantReadWriteLock.java:945)
	at org.apache.hadoop.hbase.regionserver.MetricsRegionAggregateSourceImpl.deregister(MetricsRegionAggregateSourceImpl.java:78)
	at org.apache.hadoop.hbase.regionserver.MetricsRegionSourceImpl.close(MetricsRegionSourceImpl.java:120)
	at org.apache.hadoop.hbase.regionserver.MetricsRegion.close(MetricsRegion.java:41)
	at org.apache.hadoop.hbase.regionserver.HRegion.doClose(HRegion.java:1500)
	at org.apache.hadoop.hbase.regionserver.HRegion.close(HRegion.java:1344)
	- locked <0x00000007ff878190> (a java.lang.Object)
	at org.apache.hadoop.hbase.regionserver.handler.CloseRegionHandler.process(CloseRegionHandler.java:102)
	at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:103)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
	at java.lang.Thread.run(Thread.java:744)

They are trying to MetricsRegionAggregateSourceImpl.deregister. They want to get a write lock on this classes local ReentrantReadWriteLock while holding MetricsRegionSourceImpl's readWriteLock write lock.

Then, elsewhere the JmxCacheBuster is running trying to get metrics with above locks held in reverse:

"HBase-Metrics2-1" daemon prio=5 tid=0x00007ff65e14b000 nid=0x59a03 waiting on condition [0x0000000140ea5000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x00000007cade1480> (a java.util.concurrent.locks.ReentrantReadWriteLock$NonfairSync)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:834)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireShared(AbstractQueuedSynchronizer.java:964)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireShared(AbstractQueuedSynchronizer.java:1282)
	at java.util.concurrent.locks.ReentrantReadWriteLock$ReadLock.lock(ReentrantReadWriteLock.java:731)
	at org.apache.hadoop.hbase.regionserver.MetricsRegionSourceImpl.snapshot(MetricsRegionSourceImpl.java:193)
	at org.apache.hadoop.hbase.regionserver.MetricsRegionAggregateSourceImpl.getMetrics(MetricsRegionAggregateSourceImpl.java:115)
	at org.apache.hadoop.metrics2.impl.MetricsSourceAdapter.getMetrics(MetricsSourceAdapter.java:195)
	at org.apache.hadoop.metrics2.impl.MetricsSourceAdapter.updateJmxCache(MetricsSourceAdapter.java:172)
	at org.apache.hadoop.metrics2.impl.MetricsSourceAdapter.getMBeanInfo(MetricsSourceAdapter.java:151)
	at com.sun.jmx.interceptor.DefaultMBeanServerInterceptor.getNewMBeanClassName(DefaultMBeanServerInterceptor.java:333)
	at com.sun.jmx.interceptor.DefaultMBeanServerInterceptor.registerMBean(DefaultMBeanServerInterceptor.java:319)
	at com.sun.jmx.mbeanserver.JmxMBeanServer.registerMBean(JmxMBeanServer.java:522)
	at org.apache.hadoop.metrics2.util.MBeans.register(MBeans.java:57)
	at org.apache.hadoop.metrics2.impl.MetricsSourceAdapter.startMBeans(MetricsSourceAdapter.java:221)
	- locked <0x00000007e654bdc0> (a org.apache.hadoop.metrics2.impl.MetricsSourceAdapter)
	at org.apache.hadoop.metrics2.impl.MetricsSourceAdapter.start(MetricsSourceAdapter.java:96)
	at org.apache.hadoop.metrics2.impl.MetricsSystemImpl.registerSource(MetricsSystemImpl.java:245)
	- locked <0x0000000754302660> (a org.apache.hadoop.metrics2.impl.MetricsSystemImpl)
	at org.apache.hadoop.metrics2.impl.MetricsSystemImpl$1.postStart(MetricsSystemImpl.java:229)
	at sun.reflect.GeneratedMethodAccessor45.invoke(Unknown Source)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:606)
	at org.apache.hadoop.metrics2.impl.MetricsSystemImpl$3.invoke(MetricsSystemImpl.java:290)
	at com.sun.proxy.$Proxy13.postStart(Unknown Source)
	at org.apache.hadoop.metrics2.impl.MetricsSystemImpl.start(MetricsSystemImpl.java:185)
	- locked <0x0000000754302660> (a org.apache.hadoop.metrics2.impl.MetricsSystemImpl)
	at org.apache.hadoop.metrics2.impl.JmxCacheBuster$JmxCacheBusterRunnable.run(JmxCacheBuster.java:81)

Issue Links

• relates to: HBASE-14166 Per-Region metrics can be stale