[HBASE-8558] Add timeout limit for HBaseClient dataOutputStream

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.94.5, 0.94.14
• Fix Version/s: 0.94.16
• Component/s: Client
• Hadoop Flags: Reviewed
• Assignee: Liang Xie
• Reporter: wanbin
• Created: 16/May/13 15:18
• Updated: 21/Jan/14 01:31
• Resolved: 21/Dec/13 08:43

Description

I run jstack at client host. The result is below.
"hbase-tablepool-60-thread-34" daemon prio=10 tid=0x00007f1e65a48000 nid=0x5173 runnable [0x00000000579cc000]
java.lang.Thread.State: RUNNABLE
at sun.nio.ch.EPollArrayWrapper.epollWait(Native Method)
at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:210)
at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:65)
at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:69)

• locked <0x0000000758cb0780> (a sun.nio.ch.Util$2)
• locked <0x0000000758cb0770> (a java.util.Collections$UnmodifiableSet)
• locked <0x0000000758cb0548> (a sun.nio.ch.EPollSelectorImpl)
  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:80)
  at org.apache.hadoop.net.SocketIOWithTimeout$SelectorPool.select(SocketIOWithTimeout.java:336)
  at org.apache.hadoop.net.SocketIOWithTimeout.doIO(SocketIOWithTimeout.java:158)
  at org.apache.hadoop.net.SocketOutputStream.write(SocketOutputStream.java:153)
  at org.apache.hadoop.net.SocketOutputStream.write(SocketOutputStream.java:114)
  at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:65)
  at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:123)
• locked <0x0000000754e978a0> (a java.io.BufferedOutputStream)
  at java.io.DataOutputStream.flush(DataOutputStream.java:106)
  at org.apache.hadoop.hbase.ipc.HBaseClient$Connection.sendParam(HBaseClient.java:620)
• locked <0x0000000754e97880> (a java.io.DataOutputStream)
  at org.apache.hadoop.hbase.ipc.HBaseClient.call(HBaseClient.java:975)
  at org.apache.hadoop.hbase.ipc.WritableRpcEngine$Invoker.invoke(WritableRpcEngine.java:86)
  at $Proxy13.multi(Unknown Source)
  at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$3$1.call(HConnectionManager.java:1395)
  at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$3$1.call(HConnectionManager.java:1393)
  at org.apache.hadoop.hbase.client.ServerCallable.withoutRetries(ServerCallable.java:210)
  at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$3.call(HConnectionManager.java:1402)
  at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$3.call(HConnectionManager.java:1390)
  at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)
  at java.util.concurrent.FutureTask.run(FutureTask.java:138)
  at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
  at java.lang.Thread.run(Thread.java:662)

This thread have hung for one hours

Meanwhile other thread try to close connection

"IPC Client (1983049639) connection to dump002030.cm6.tbsite.net/10.246.2.30:30020 from admin" daemon prio=10 tid=0x00007f1e70674800 nid=0x3d76 waiting for monitor entry [0x000000004bc0f000]
java.lang.Thread.State: BLOCKED (on object monitor)
at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:123)

• waiting to lock <0x0000000754e978a0> (a java.io.BufferedOutputStream)
  at java.io.DataOutputStream.flush(DataOutputStream.java:106)
  at java.io.FilterOutputStream.close(FilterOutputStream.java:140)
  at org.apache.hadoop.io.IOUtils.cleanup(IOUtils.java:237)
  at org.apache.hadoop.io.IOUtils.closeStream(IOUtils.java:254)
  at org.apache.hadoop.hbase.ipc.HBaseClient$Connection.close(HBaseClient.java:715)
• locked <0x0000000754e7b818> (a org.apache.hadoop.hbase.ipc.HBaseClient$Connection)
  at org.apache.hadoop.hbase.ipc.HBaseClient$Connection.run(HBaseClient.java:587)

dump002030.cm6.tbsite.net is dead regionserver.

I read hbase sourececode, discover connection.out doesn't set timeout

this.out = new DataOutputStream
(new BufferedOutputStream(NetUtils.getOutputStream(socket)));

I see this mean epoll_wait will block indefinitely.