JDBC read stuck if RAC node goes down

Question

We did several tests with Java applications against our RAC DB and face a hanging application if we power off the RAC node that executes the current (long) running query.
We can see that the application receives HA-events via UCP:
2015-01-22 13:02:11 | r-thread-1 | WARN  | o.ucp.jdbc.oracle.ONSDatabaseFailoverEvent    | NO timezone in HA event
However, the application started a query before and the query is not aborted with an exception. A Thread dump after about 7 minutes shows that the application is hanging in a socket read call:
"pool-1-thread-1" #32 prio=5 os_prio=0 tid=0x00007fedf45b2000 nid=0xbc4 runnable [0x00007fee00cd3000]
   java.lang.Thread.State: RUNNABLE
    at java.net.SocketInputStream.socketRead0(Native Method)
    at java.net.SocketInputStream.read(SocketInputStream.java:150)
    at java.net.SocketInputStream.read(SocketInputStream.java:121)
    at oracle.net.ns.Packet.receive(Packet.java:283)
    at oracle.net.ns.DataPacket.receive(DataPacket.java:103)
    at oracle.net.ns.NetInputStream.getNextPacket(NetInputStream.java:230)
    at oracle.net.ns.NetInputStream.read(NetInputStream.java:175)
    at oracle.net.ns.NetInputStream.read(NetInputStream.java:100)
    at oracle.net.ns.NetInputStream.read(NetInputStream.java:85)
    at oracle.jdbc.driver.T4CSocketInputStreamWrapper.readNextPacket(T4CSocketInputStreamWrapper.java:123)
    at oracle.jdbc.driver.T4CSocketInputStreamWrapper.read(T4CSocketInputStreamWrapper.java:79)
    at oracle.jdbc.driver.T4CMAREngine.unmarshalUB1(T4CMAREngine.java:1122)
    at oracle.jdbc.driver.T4CMAREngine.unmarshalSB1(T4CMAREngine.java:1099)
    at oracle.jdbc.driver.T4CTTIfun.receive(T4CTTIfun.java:288)
    at oracle.jdbc.driver.T4CTTIfun.doRPC(T4CTTIfun.java:191)
    at oracle.jdbc.driver.T4C8Oall.doOALL(T4C8Oall.java:523)
    at oracle.jdbc.driver.T4CPreparedStatement.doOall8(T4CPreparedStatement.java:207)
    at oracle.jdbc.driver.T4CPreparedStatement.executeForDescribe(T4CPreparedStatement.java:863)
    at oracle.jdbc.driver.OracleStatement.executeMaybeDescribe(OracleStatement.java:1153)
    at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1275)
    at oracle.jdbc.driver.OraclePreparedStatement.executeInternal(OraclePreparedStatement.java:3576)
    at oracle.jdbc.driver.OraclePreparedStatement.executeQuery(OraclePreparedStatement.java:3620)
    - locked <0x00000000c0ddcb20> (a oracle.jdbc.driver.T4CConnection)
    at oracle.jdbc.driver.OraclePreparedStatementWrapper.executeQuery(OraclePreparedStatementWrapper.java:1491)
    at org.springframework.jdbc.core.JdbcTemplate$1.doInPreparedStatement(JdbcTemplate.java:703)
The expected behaviour would be that a running query is aborted with an exception. (BTW: This happens if the service is taken down with "shutdown immediate". All ok for this case.)
We consider to implement custom ONS listeners [1], but we actually expect that UCP would handle such situations or lets us register strategies/callbacks for certain events.
Our config:
Oracle Enterprise 11.2.0.4.0 with RAC
ons.jar 12.1.0.1
ojdbc6.jar 11.2.0.2
ucp.jar 12.1.0.1
Server JRE 1.8.0_25
Any hints appreciated.
[1] http://docs.oracle.com/cd/E11882_01/java.112/e16548/apxracfan.htm#JJDBC28945

Answer 1

You're concept isn't right:
http://docs.oracle.com/cd/E11882_01/server.112/e25494/restart.htm#ADMIN13178
Overview of Fast Application Notification
FAN is a notification mechanism that Oracle Restart can use to notify other processes about configuration changes that include service status changes, such as UP or DOWN events. FAN provides the ability to immediately terminate inflight transaction when an instance or server fails. Integrated Oracle clients receive the events and respond. Applications can respond either by propagating the error to the user or by resubmitting the transactions and masking the error from the application user. When a DOWN event occurs, integrated clients immediately clean up connections to the terminated database. When an UP event occurs, the clients create new connections to the new primary database instance.
Also, take a look at these docs: http://docs.oracle.com/cd/E11882_01/java.112/e12265/rac.htm#JJUCP08100 ; and https://support.oracle.com/epmos/faces/DocumentDisplay?_afrLoop=890204623685515&id=566573.1&_afrWindowMode=0&_adf.ctrl-s…
And make a test, execute  a query that took about 1 minute and after you executed, just power down the node where it is executing, to see if it will retrieve the results.
Regards.