Cluster link and multicast

thumbnail
1498384, modified 15 Years ago. Liferay Legend Posts: 3089 Join Date: 10/28/08 Recent Posts
Hi all!

I've searching for a while to know what is the best cluster configuration.

I know cluster link enables clustering for indexes but, does it enable clustering for something else?

Looking into portal.properties I've found this properties for configuring multicast:


 
    multicast.group.address["cluster-link-control"]=239.0.0.12
    multicast.group.port["cluster-link-control"]=23301

    #
    # See the properties "cluster.link.channel.properties.transport.0" and
    # "cluster.link.channel.system.properties".
    #
    multicast.group.address["cluster-link-udp"]=239.0.0.13
    multicast.group.port["cluster-link-udp"]=23302

    #
    # See the property "cluster.link.channel.system.properties".
    #
    multicast.group.address["cluster-link-mping"]=239.0.0.14
    multicast.group.port["cluster-link-mping"]=23303

    #
    # See the properties "net.sf.ehcache.configurationResourceName" and
    # "net.sf.ehcache.configurationResourceName.peerProviderProperties".
    #
    multicast.group.address["hibernate"]=239.0.0.15
    multicast.group.port["hibernate"]=23304

    #
    # See the properties "ehcache.multi.vm.config.location" and
    # "ehcache.multi.vm.config.location.peerProviderProperties".
    #
    multicast.group.address["multi-vm"]=239.0.0.16
    multicast.group.port["multi-vm"]=23305


If I enable cluster link, isn't enabled by default the "hibernate" clustering and "multi-vm" clustering, or do I have to enable by myself if I want total clustering?

Thanks very much!
thumbnail
4747417, modified 15 Years ago. Junior Member Posts: 39 Join Date: 3/22/10 Recent Posts
Hi Juan,

ClusterLink only clusters the indexes.


You will also need to cluster your Ehcache, have a shared database, and will probably need to share a document library repository.

To cluster your Ehcache.
Put these properties in your portal-ext.properties to do so:

    
ehcache.multi.vm.config.location=/ehcache/liferay-multi-vm-clustered.xml
net.sf.ehcache.configurationResourceName=/ehcache/hibernate-clustered.xml


Hope this helps.
thumbnail
1498384, modified 15 Years ago. Liferay Legend Posts: 3089 Join Date: 10/28/08 Recent Posts
Thanks very much Christopher! Totally clear now.
thumbnail
Bla Bla, modified 15 Years ago. Junior Member Posts: 79 Join Date: 12/22/10 Recent Posts
Hi,

I have activated the multicast on my servers and added the following properties in the portal-ext.properties.

ehcache.multi.vm.config.location=/ehcache/liferay-multi-vm-clustered.xml
net.sf.ehcache.configurationResourceName=/ehcache/hibernate-clustered.xml


But when I alter a WCM content on my fist liferay instance, I cannot see it right away on the second. I have to wait (at least that's my guess) that hibernate cache expires to see the altered content on the second liferay.

My two liferay instances are behind an apache server each, and each apache server is behinf a physical load balancer.

Any guess ? Anything more to configure ?

Something more : I found 5 addresses in the portal.properties file related to the multicast / cluster thing : 233.0.0.1 to 233.0.0.5.

When I try to ping those addresses from one of my two liferay server, each IP address seems to respond correctly (1 packet send, 1 packet received, 1 duplicate => multicast appears to work) BUT for the 233.0.0.3 address which do not respond to the ping command.

What is this IP address for (that's the "multicast.group.address["cluster-link-mping"]" property) ? Could it be related to my problem ?

Thank you VERY much for any clue !

Mathieu.
thumbnail
Mathieu Hicauber, modified 15 Years ago. Junior Member Posts: 79 Join Date: 12/22/10 Recent Posts
Problem solved for me. I answer mysefl, in case it helps others.

The problem was due to a bad configuration of the /etc/hosts file. In clustered mode, RMI calls are made that use a getHostName function; this function reliez on this file to get the IP adress of the machine.
In my case, the configuration was set this way :
127.0.1.1 NAME_OF_THE_MACHINE

While this will work for most of the case, for RMI communication the IP address must be the "real" IP address of the machine.

Finally it works like a charm :-)
Артем Карузин, modified 6 Years ago. New Member Posts: 2 Join Date: 10/30/15 Recent Posts
Mathieu Hicauber:

Problem solved for me. I answer mysefl, in case it helps others.

The problem was due to a bad configuration of the /etc/hosts file. In clustered mode, RMI calls are made that use a getHostName function; this function reliez on this file to get the IP adress of the machine.
In my case, the configuration was set this way :
127.0.1.1 NAME_OF_THE_MACHINE

While this will work for most of the case, for RMI communication the IP address must be the "real" IP address of the machine.

Finally it works like a charm :-)

Thank you! very helpful for me!
thumbnail
8254193, modified 15 Years ago. Regular Member Posts: 197 Join Date: 4/15/11 Recent Posts
Hi ,

I have done clustering using two nodes in Linux and Windows both.But When i am adding one more node,In Linux it is working fine but in windows after starting 3rd node stops.Is this problem because of Multicasting?

How to configure Multicasting in windows environment?


Thanks in advance
thumbnail
Stephan Huber, modified 14 Years ago. Junior Member Posts: 28 Join Date: 3/27/09 Recent Posts
Is there a reason why there are five different Multicast IP addresses used in the portal.properties?

As the five Multicast ports are already different shouldn't there be one Multicast IP address enough?

Thanks!