[HBASE-14926] Hung ThriftServer; no timeout on read from client; if client crashes, worker thread gets stuck reading

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.2.0, 1.1.2, 1.3.0, 1.0.3, 0.98.16, 2.0.0
• Fix Version/s: 1.2.0, 1.3.0, 0.98.17, 2.0.0
• Component/s: Thrift
• Hadoop Flags: Reviewed
• Release Note: Adds a timeout to server read from clients. Adds new configs hbase.thrift.server.socket.read.timeout for setting read timeout on server socket in milliseconds. Default is 60000
• Assignee: Michael Stack
• Reporter: Michael Stack
• Created: 04/Dec/15 06:25
• Updated: 05/Dec/15 13:11
• Resolved: 04/Dec/15 22:06

Description

Thrift server is hung. All worker threads are doing this:

"thrift-worker-0" daemon prio=10 tid=0x00007f0bb95c2800 nid=0xf6a7 runnable [0x00007f0b956e0000]
   java.lang.Thread.State: RUNNABLE
        at java.net.SocketInputStream.socketRead0(Native Method)
        at java.net.SocketInputStream.read(SocketInputStream.java:152)
        at java.net.SocketInputStream.read(SocketInputStream.java:122)
        at java.io.BufferedInputStream.fill(BufferedInputStream.java:235)
        at java.io.BufferedInputStream.read1(BufferedInputStream.java:275)
        at java.io.BufferedInputStream.read(BufferedInputStream.java:334)
        - locked <0x000000066d859490> (a java.io.BufferedInputStream)
        at org.apache.thrift.transport.TIOStreamTransport.read(TIOStreamTransport.java:127)
        at org.apache.thrift.transport.TTransport.readAll(TTransport.java:84)
        at org.apache.thrift.transport.TFramedTransport.readFrame(TFramedTransport.java:129)
        at org.apache.thrift.transport.TFramedTransport.read(TFramedTransport.java:101)
        at org.apache.thrift.transport.TTransport.readAll(TTransport.java:84)
        at org.apache.thrift.protocol.TCompactProtocol.readByte(TCompactProtocol.java:601)
        at org.apache.thrift.protocol.TCompactProtocol.readMessageBegin(TCompactProtocol.java:470)
        at org.apache.thrift.TBaseProcessor.process(TBaseProcessor.java:27)
        at org.apache.hadoop.hbase.thrift.TBoundedThreadPoolServer$ClientConnnection.run(TBoundedThreadPoolServer.java:289)
        at org.apache.hadoop.hbase.thrift.CallQueue$Call.run(CallQueue.java:64)
        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:745)

They never recover.

I don't have client side logs.

We've been here before: HBASE-4967 "connected client thrift sockets should have a server side read timeout" but this patch only got applied to fb branch (and thrift has changed since then).