APAR status |
Closed as program error.
| Error description
They are running Web Methdods as a Java client on Solaris,
connecting to AS 3.5.5 statefull session bean on another
Solaris machine.
This statefull session bean communicates to a stateless
session bean on another remote 3.5.5 install on AIX. The
remote tier is actually two 3.5.5 nodes in the same domain, 01
and 02. 01 is the primary node and contains a model and
cloned application server. The application server clone is sit
ting on 02. When the clone on 02 is stopped, the application
does not connect to the clone still running.
.
Actual Exception:
=================
[1/15/02 13:29:34:329 CST] 21b220 ORBRas X
com.ibm.CORBA.iiop.IIOPConnection send Thread-0 The following
exception was logged
java.io.InterruptedIOException:Actual Exception:=================[1/15/02 13:29:34:329 CST], 21b220 ORBRas Xcom.ibm.CORBA.iiop.IIOPConnection send Thread-0 The followingexception was logged
at java.net.SocketOutputStream.socketWrite(Native Method)
at java.net.SocketOutputStream.write(Unknown Source)
at
com.ibm.CORBA.iiop.ClientRequestImpl.writeTo(ClientRequestImpl.j
ava:336)
at
com.ibm.CORBA.iiop.IIOPConnection.send(IIOPConnection.java:1352)
at
com.ibm.CORBA.iiop.ClientRequestImpl.invoke(ClientRequestImpl.ja
va:508)
at com.ibm.CORBA.iiop.Cli
.
If the client bean were running on AIX, it does recover. On
Solaris, it does not. java.io.InterruptedIOException:at java.net.SocketOutputStream.socketWrite(Native Method)at java.net.SocketOutputStream.write(Unknown Source)atcom.ibm.CORBA.iiop.ClientRequestImpl.writeTo(ClientRequestImpl.java:336)atcom.ibm.CORBA.iiop.IIOPConnection.send(IIOPConnection.java:1352)atcom.ibm.CORBA.iiop.ClientRequestImpl.invoke(ClientRequestImpl.java:508)at com.ibm.CORBA.iiop.Cli.If the client bean were running on AIX, it does recover. OnSolaris, it does not. Local fixProblem summary
****************************************************************
* USERS AFFECTED: WebSphere Application Server users of *
* enterprise applications. *
****************************************************************
* PROBLEM DESCRIPTION: Enterprise application hangs. *
****************************************************************
* RECOMMENDATION: *
****************************************************************
Enterprise application hangs when it tries to use RMI-IIOP to
send data across the wire. This hang seems to be happening on
the client side. Problem conclusion
This defect was created to workaround a JDK bug. When ORB's
IIOPConnection.send() tries to send data onto the wire, JDK
intermittently throw an InterruptedIOException back
IIOPConnection.send() and eventually cause a hang on the ORB
layer. To avoid this InterruptedIOException, ORB explicitly
re-tries five times when this exception were thrown from JDK.
However, if JDK continuously throw the same exception or
IOException after 5 retries, ORB will be hang again. To make
ORB more robust, ORB's reader thread should be adjusted to
avoid the potential dead lock in ORB's layer. Temporary fixComments
APAR information | APAR number | PQ61243 | Reported component name | WAS ADVANCED AI | Reported component ID | 5648C8400 | Reported release | 350 | Status | CLOSED PER | PE | NoPE | HIPER | NoHIPER | Submitted date | 2002-05-16 | Closed date | 2002-06-17 | Last modified date | 2002-11-18 |
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:APAR is sysrouted FROM one or more of the following:
Modules/Macros APAR is sysrouted TO one or more of the following:Modules/Macros
|
Fix information |
Fixed component name | WAS ADVANCED AI | Fixed component ID | 5648C8400 |
Applicable component levels | R350 PSY | UP |
|