add a retry to hotrod operations

From an email from Galder:

Hi Mircea,

I'm playing around with Hot Rod client/server for JUDCon demo and I've
found this. The demo consists of:

1. Start a server
2. Connect with client and do a put
3. Make several get calls on that server
4. Start a new node, clustered with replication and state transfer
on.
5. Make several calls, the calls are being load balanced in round
robin fashion.
6. Shutdown the first node and make a get call. This throws:

Exception in thread "main" HotRodServerException

{messageId=0, errorStatusCode=0}

org.infinispan.client.hotrod.exceptions.TransportException:
java.io.IOException: Broken pipe
at
org.infinispan.client.hotrod.impl.transport.VHelper.writeVLong(VHelper.java:48)
at
org.infinispan.client.hotrod.impl.transport.tcp.TcpTransport.writeVLong(TcpTransport.java:57)
at
org.infinispan.client.hotrod.impl.protocol.HotRodOperationsHelper.writeHeader(HotRodOperationsHelper.java:30)
at
org.infinispan.client.hotrod.impl.protocol.HotRodOperationsImpl.sendKeyOperation(HotRodOperationsImpl.java:256)
at
org.infinispan.client.hotrod.impl.protocol.HotRodOperationsImpl.get(HotRodOperationsImpl.java:39)
at
org.infinispan.client.hotrod.impl.RemoteCacheImpl.get(RemoteCacheImpl.java:276)
at replication.GetOnReadLoop$.main(GetOnReadLoop.scala:19)
at replication.GetOnReadLoop.main(GetOnReadLoop.scala)
Caused by: java.io.IOException: Broken pipe
at sun.nio.ch.FileDispatcher.write0(Native Method)
at sun.nio.ch.SocketDispatcher.write(SocketDispatcher.java:29)
at sun.nio.ch.IOUtil.writeFromNativeBuffer(IOUtil.java:104)
at sun.nio.ch.IOUtil.write(IOUtil.java:75)
at sun.nio.ch.SocketChannelImpl.write(SocketChannelImpl.java:334)
at java.nio.channels.Channels.write(Channels.java:60)
at java.nio.channels.Channels.access$000(Channels.java:47)
at java.nio.channels.Channels$1.write(Channels.java:134)
at java.io.OutputStream.write(OutputStream.java:58)
at java.nio.channels.Channels$1.write(Channels.java:115)
at
org.infinispan.io.UnsignedNumeric.writeUnsignedLong(UnsignedNumeric.java:134)
at
org.infinispan.client.hotrod.impl.transport.VHelper.writeVLong(VHelper.java:46)
... 7 more

Shouldn't have this been swallowed by the client and shouldn't have it
tried on the other node?

This happens if a server shuts down. After a failure the connection is cleaned.
A workaround for this should be "testOnBorrow" which would first test the connection when fetching it from the server.