How to achieve session replication among jboss clusters in jboss As 7.1

[chaitu1011]

Hi,

I have made two standalone jboss 7 instances, and am able to load balance them using mod_jk.

i want to implement session replication in my app.

i have searched the internet and found that adding <distributable/> tag to the web.xml would enable the session replication.

i have tried it, but its not working.

for instance. if i start my both  jobss instace and run my application, and shutdown the instance where session is created, the second node(instance) creating a new session, thus session data is lost , i again have to login to my application.

could any one please explain me the configaratuion thats required in jboss 7for session replication

any help is appriciated

Thanks

[chaitu1011]

any one guys ??

any infomation is highly appreciated.

thanks

[rhusar]

I have made two standalone jboss 7 instances, and am able to load balance them using mod_jk.

Just make sure you are running standalone-ha.xml or standalone-full-ha.xml profiles that include clustering.

Also make sure that you have sticky sessions on, and stick sessions force off.

Check the log that the logs show that multiple nodes joined in.

i have searched the internet and found that adding <distributable/> tag to the web.xml would enable the session replication.

Yes that is enough. But you need to make sure that after each change to the session you call on HttpSession.setAttribute(xyz) to tell clustering that the attribute has changed and should replicate.

Rado

[chaitu1011]

Hi Rado

Thanks a lot for your reply,

so if i run jboss with standalone-ha.xml, dont i need to create multiple instances of jboss ??

could be please elaborate on that ?

and alo i have sticky session = true in my workers.properties file .

Thanks

[rhusar]

so if i run jboss with standalone-ha.xml, dont i need to create multiple instances of jboss ??

Well, if you want reasonable clustering you always need at least 2 nodes, right? ;-) So on each machine you need to unzip and start separately. So yes, you create multiple instances.

Make sure you bind (-b switch) to "public" IP addresses, meaning the instances can see each other and communicate.

and alo i have sticky session = true in my workers.properties file .

Yap, that's good when your application is stateful.

Rado

[chaitu1011]

Hi Rado,

Thanks, i have configured the changes as you said. but still unable to acheive session replication .

I am using mod_jk for load balancing by the way.

below is my worker.properties file

# Define list of workers that will be used

# for mapping requests

worker.list= loadbalancer,status,node1,node2

# Define Node1

# modify the host as your host IP or DNS name.

worker.node1.port=8009

worker.node1.host=127.0.0.1

worker.node1.type=ajp13

worker.node1.lbfactor=1

#worker.node1.prepost_timeout=10000

#worker.node1.connect_timeout=10000

worker.node1.connection_pool_size=10

# Define Node2

# modify the host as your host IP or DNS name.

worker.node2.port=8010

worker.node2.host=192.168.16.228

worker.node2.type=ajp13

worker.node2.lbfactor=1

#worker.node2.prepost_timeout=10000

#worker.node2.connect_timeout=10000

worker.node2.connection_pool_size=10 

# Load-balancing behaviour

worker.loadbalancer.balance_workers=node1,node2

worker.loadbalancer.type=lb

worker.loadbalancer.sticky_session=force

  sticky-session-force=false

# Status worker for managing load balancer

worker.status.type=status

and things i have configured in standalone-ha.xml

xml1

<subsystem xmlns="urn:jboss:domain:web:1.1" default-virtual-server="default-host" native="false" instance-id="node1">

xml2

<subsystem xmlns="urn:jboss:domain:web:1.1" default-virtual-server="default-host" native="false" instance-id="node2">

i can see in the logs that the clustering is happeneing.

please let me know if i am missing any configarations

Thanks a ton for your help .. 

[chaitu1011]

any one guys ...

[rhusar]

Does that configuration even work?

worker.list= loadbalancer,status,node1,node2

should be

worker.list=loadbalancer,status

worker.node2.port=8010

Doing port offset like this is not very good. Bind to different IP adresses (different like 10.0.0.1 and 10.0.0.2) and dont use port offset.

worker.node2.connection_pool_size=10 

You can leave the default value here IMHO, so you dont have to specify it.

worker.loadbalancer.sticky_session=force

AFAIR it should be worker.loadbalancer.sticky_session=True

The instance-id are correct.

[chaitu1011]

Hi Rado,

Thanks a lot for your replies

i have tried by not using offset.

i am binding node 1 with -b 127.0.0.1(localhost)

and node 2 with -b 192.168.16.228

i am able to load balance the instances but not able to replicate the session

say i start node 1 like this

standalone.bat -b 127.0.0.1 and node 2 like standalone.bat -b 192.168.16.228

if i try to access my app, either one of the nodes will serve the request. if say, node 2 servered the request and if i kill node 2 new session is being created on node1 but the session info is lost

i dunno what conig i am missing

by the way, i am using apache 2.2 as front end with mod_jk and jboss 7.1.1

thanks

[rhusar]

Can you check the log and see that the server really formed? First the joining node outputs the number of cluster nodes

17:39:27,931 INFO  [org.jboss.as.clustering.impl.CoreGroupCommunicationService.web] (MSC service thread 1-1) JBAS010206: Number of cluster members: 2

and then you should see the cluster view that contains both nodes

17:39:28,382 INFO  [org.infinispan.remoting.transport.jgroups.JGroupsTransport] (pool-15-thread-1) ISPN000094: Received new cluster view: [rhusar1/web|1] [rhusar1/web, rhusar2/web]

this looks ok

by the way, i am using apache 2.2 as front end with mod_jk and jboss 7.1.1

Maybe you are testing it wrong? :-) Make sure you are accessing the load-balancer, not the sessions directly.

[chaitu1011]

Hi again Rado,

if i run standalone -b 127.0.0.1 and standalone -b 192.168.16.228

it doesnt show me that number of clustered members =2

but if i run like standalone -b 192.xxx.xx.xx

and

standalone -b 192.168.16.228

it shows me num of clustered nodes=2

[ctomc]

Hi,

you should be running cluster configuration of AS

standalone-ha.xml or standalone-full-ha.xml are the ones

so for your use case you need to do something like this:

standalone -c standalone-ha.xml -b 192.168..xxx

might be that you will need to apply same configration changes that you did to original standalone.xml

--

tomaz