Hide Forgot
Date of First Response: 2010-04-20 14:56:23 project_key: JBEPP I have 2 nodes with GateIn portal and one Apache loadbalancer with mod_jk and sticky session enabled. Steps to simulate: 1) Run loadbalancer 2) Run node1 with -Dexo.profiles=cluster 3) Run node2 with -Dexo.profiles=cluster 4) Login as root in loadbalancer and go to portalnavigation -> Create new portal -> Do all steps to create new portal called "NewPortal" with the wizard. 5) Go to http://loadbalancer:8080/portal/private/NewPortal and everything is fine 6) Kill the active node of loadbalancer, which served the current HTTP session 7) Go to http://loadbalancer:8080/portal/private/NewPortal. Now I am not able to see new portal - only blank screen and exception in server log. I am also not seeing new portal in portalnavigation . Logout of user and login again does not help with it. Only restart helps ( I can see new portal on node2 after restart.)
Attachment: Added: NullPointerException-cluster-createPortal.txt
I am able to simulate the issue also without usage of loadbalancer. 1) Start node1 2) Start node2 3) Login as root in node1 and create new portal here. Everything is fine. 4) Go to node2 and verify that new portal can't be seen here in portalnavigation and blanks screen (with NPE in server log) appears when trying to access it -> ERROR 5) Restart node2. New portal can be seen now.
I think that this scenario works before in EPP5 ER4. But I am not sure at 100%...
Link: Added: This issue is related to GTNPORTAL-1104
I reproduced this scenario with current portal trunk and when going to node2, everything went fine, can you check on your side ?
I am also not able to reproduce with latest trunk from today (revision 2716 - after fix of GTNPORTAL-1109). I am seeing portal on node2 directly after creation and also looks good with loadbalancer. I tried various scenarios to test but looks that portal creation, deletion works fine in 2-nodes cluster now.
I also tried and was not able to reproduce