Red Hat Bugzilla – Bug 197805
jvm multicast binding on RHEL4 does not works
Last modified: 2007-11-30 17:07:26 EST
Description of problem:
We use Bea Weblogic 8.1SP5 with Jrockit and Sun JVM 1.4.2 on RHEL3. We upgraded
to operating system to RHEL4, and found that the interface binding for
multicast communication does not bind to the correct interface, it is always
using the default interface. We tried it from a sample program, and found that
the error is in the JVMs(on RHEL4 kernel???).
This is a critical error for us, because our application server cluster use
multicast for cluster communication, so on RHEL4 it does not works.
No error msg. It seems like this BUG:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4082533 but it is solved.
Version-Release number of selected component (if applicable):
OS and Kernel Version: Linux achilles1 2.6.9-34.0.1.ELsmp #1 SMP Wed May 17
16:59:36 EDT 2006 x86_64 x86_64 x86_64 GNU/Linux
List steps to reproduce the problem:
socket = new MulticastSocket(i);
if(s2 != null)
InetAddress inetaddress = InetAddress.getByName(s2);
System.out.println("Using interface at " + inetaddress.getHostAddress());
always bind to the default interface
bind to the selected interface
We restored the whole Bea and JVM home from backup, so the only difference was
is the RHEL3->RHEL4 change. Something could be wrong with the JVM-RHEL4
multicast interface binding.
We found the real problem:
The setInterface(InetAddress inf) method does not works on RHEL4. If we use it
on RHEL3 sets the output network interface, but on RHEL4 it is always using the
default interface for sending multicast packets even we set other interface
before the send. The Weblogic multicast tester (and cluster..) using
setInterface(InetAddress inf) method to set the network interface; so the
cluster does not works on RHEL4 if you are not using the default network
interface for WLS cluster interconnect.
We asked Bea and Sun to revoke their RHEL4 platform certification for Java and
Weblogic until this type of problems are exists.
We found a workaround:-Djava.net.preferIPv4Stack=true
There should be a big problem in the IPV6 multicast stack on RHEL4..
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux maintenance release. Product Management has requested
further review of this request by Red Hat Engineering, for potential
inclusion in a Red Hat Enterprise Linux Update release for currently deployed
products. This request is not yet committed for inclusion in an Update
Could you please update the status uf this Bug? Are there any decision?
(In reply to comment #7)
> Could you please update the status uf this Bug? Are there any decision?
We're working on this for RHEL-4.5.
Georges, are you aware of this bug? Is it fixed in 1.4.2_11 R26.4.0-63?
Could you please update the status uf this Bug? Is is not VM specific, it is a
kernel bug. What RHEL4-5 mean? RHEL4 update 5 or something else?
(In reply to comment #14)
> Could you please update the status uf this Bug? Is is not VM specific, it is a
> kernel bug.
If you have a standalone test case that reproduces the problem and demonstrates
that it is a kernel problem, then please file a separate bug report for it.
> What RHEL4-5 mean? RHEL4 update 5 or something else?
Yes, RHEL-4 Update 5.
If this is just a case of JRockit not being updated with a Sun bug fix, then the
upcoming RHEL-4.5 JRockit erratum should contain the fix, so I'm leaving this
open until then. I recommend re-testing at that point.
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.