BE-MM fails to start as error “java.lang.RuntimeException: com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)” is thrown.

BE-MM fails to start as error “java.lang.RuntimeException: com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)” is thrown.

book

Article ID: KB0091196

calendar_today

Updated On:

Products Versions
TIBCO BusinessEvents Enterprise Edition -
Not Applicable -

Description

Resolution:
Description:
============
BE-MM fails to start as error “java.lang.RuntimeException: com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)” is thrown.


Environment:
============
TIBCO BusinessEvents 5.1.0
Red Hat Linux 9.0


Symptoms:
============
When running BE-MM.exe the following error is thrown.

fatal [main] - [container.standalone] com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)
java.lang.RuntimeException: com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)



Cause:
=======
Check the following:

1). MM-CDD file and ensure that the properties:
“be.engine.cluster.as.listen.url” and “be.engine.cluster.as.discover.url” have been correctly defined and there are no syntax errors.

2). Check /etc/hosts file.
If the desired interface is not listed in the hosts file, then the first available interface in the file will be picked up. If the first interface in the hosts file is a loopback interface (127.0.0.1) then it will not work.

In this particular case, 127.0.0.1 was pointing to the machine name which is incorrect. After updating the hosts file so that the machine IP address was referencing the machine name, the error was resolved.

3). If the above does not resolve the issue check if multicasting is disabled on the box. If this is the case, they have to use unicast.

Configuration Unicast:
Every PU must have its own listen URL and all PUs should have the same discovery URL. So discovery URL can be defined as a Cluster level property while listen URL needs to be a PU level property. The discovery URL (if using TCP Unicast) needs to be a sub-set of all the listen URLs of all the PUs. At least one of the PUs whose listen URL is part of the discovery URL needs to be up and running.

be.engine.cluster.as.discover.url -> cluster level includes all members
be.engine.cluster.as.listen.url -> set in PU of CDD with one of the ports defined in discover.url


Resolution:
============
127.0.0.1 was pointing to the machine name which is incorrect. Update /etc/hosts file where machine IP address was referencing the machine name and not the localhost entry.

Issue/Introduction

BE-MM fails to start as error “java.lang.RuntimeException: com.tibco.as.space.ASException: SYS_ERROR (multicast_error - (8) grp_iface not a valid multicast interface)” is thrown.