powered by Jive Software

Hazelcast Clustering

Hello Friends,

I have recently deployed Openfire 3.8.2 with Hazelcast Clustering 1.0.6. After enabling the cluster I am getting error on openfire console.

Starting Hazelcast Clustering Plugin

30 Jan, 2014 6:19:12 PM com.hazelcast.config.ClasspathXmlConfig

INFO: Configuring Hazelcast from ‘hazelcast-cache-config.xml’.

30 Jan, 2014 6:19:12 PM com.hazelcast.impl.AddressPicker

INFO: Prefer IPv4 stack is true.

30 Jan, 2014 6:19:12 PM com.hazelcast.impl.AddressPicker

INFO: Picked Address[172.16.0.73]:5701, using socket ServerSocket[addr=/0.0.0.0,localport=5701], bind any local is true

30 Jan, 2014 6:19:12 PM com.hazelcast.system

INFO: [172.16.0.73]:5701 [openfire] Hazelcast Community Edition 2.5.1 (20130427) starting at Address[172.16.0.73]:5701

30 Jan, 2014 6:19:12 PM com.hazelcast.system

INFO: [172.16.0.73]:5701 [openfire] Copyright © 2008-2013 Hazelcast.com

30 Jan, 2014 6:19:12 PM com.hazelcast.impl.LifecycleServiceImpl

INFO: [172.16.0.73]:5701 [openfire] Address[172.16.0.73]:5701 is STARTING

30 Jan, 2014 6:19:15 PM com.hazelcast.impl.MulticastJoiner

INFO: [172.16.0.73]:5701 [openfire]

Members [1] {

Member [172.16.0.73]:5701 this

}

30 Jan, 2014 6:19:15 PM com.hazelcast.impl.LifecycleServiceImpl

INFO: [172.16.0.73]:5701 [openfire] Address[172.16.0.73]:5701 is STARTED

30 Jan, 2014 6:19:15 PM com.hazelcast.impl.PartitionManager

INFO: [172.16.0.73]:5701 [openfire] Initializing cluster partition table first arrangement…

RMS

I am also facing similar problem with the latest update. Is there any solution for this?

where is the error ? i dont see an error msg in your log.

if you have more members, maybe you have a problem with configuration.

Hi Daniel,

We are also facing the same issue…and there is indeed no error log generated.

The problem is that we have a configuration with 2 nodes (configured in unicast)…that was working perfectly under a OF v3.8.x + Hazelcast plugin v1.0…And since we have updated everything with OF v3.9.1 + Hazelcast plug-in v1.2…the nodes cannot see each others.

Configuration is the following (ucxmpp01 and ucxmpp02 are obviously configured in the hosts file):

5701


224.2.2.3
54327


ucxmpp01
ucxmpp02


Do you have any clue that could help us to troubleshoot/fix it ?

Regards

Christophe

Hi Christophe -

You might want to do a quick diff/compare of your Hazelcast config file with the latest version shipped with the plugin. If I recall correctly, a few of the policy configuration constants changed in Hazelcast version 3.0. You should be able to make the same changes in your custom configuration file.

Hope that helps. If I get an opportunity I will try to dig up the specific changes and update the plugin’s README file accordingly. Please let us know how it goes.

Regards,

Tom

I have solved this problem by adding interface section. I was also having the similar config as described above (with hostnames) and started not joining the cluster after I upgrade to 3.9.1. I think earlier versions if we just provide hostnames in section it used to join but now need to provide interfaces to get connected. The below config works with 3.9.1

ip_of_node_1 ip_of_node_2 IP_ranges_covers both the node