We are running a two node Resin 4.0.12 cluster behind Barracuda 440 Load 
Balancers. I start one server (server-a) in the cluster and log into an app 
(e.g. generate a session). I then start the second server (server-b) in the 
cluster (wait for it to completely start) and then shut down server-a. I then 
click a link in the application I previously logged into at which point my 
application cannot find my session and I have to log in again. I can run the 
same steps (reversing the servers) and generate the behavior again and again.

Is this a valid test case for session replication? I would expect that the 
session information is transmitted across the cluster and when I click the link 
I would be seamlessly directed to server-b which would have my session 
information.

Is anyone else running Resin 4.0.12 in a cluster behind hardware load 
balancers? If so, have you experienced and/or are you willing to test this 
behavior?

Key Resin config entries:
<web-app-default>
...
<session-config use-persistent-store="true" always-save-session="true" />
...
</web-app-default>

<cluster id="app-tier">
...
<server-default>
...
<host port="80"/>
...
</server-default>
...
<server id="server-a" address="10.1.1.2" port="6800"/> 
<server id="server-b" address="10.1.1.3" port="6800"/>
<persistent-store type="cluster" />
...
</cluster>


thanks
matt

_______________________________________________________
Matthew Serrano - Director of Engineering
Pearson Knowledge Technologies - Ordinate Corporation
299 S. California Ave Suite 300
Palo Alto, CA 94306
650 470 3600
650 470 3533 (direct)
mserr...@pearsonkt.com
www.ordinate.com
www.pearson.com

_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest

Reply via email to