Hi All,
In one of our project we have two MWS server say node1 and node2, both are in cluster. We are using apache as reverse proxy and load balancer as well. Everything is going fine I am able to access MWS via Apache webserver and it is doing load balancing as well but issue is in load balancer when one node is down then balancer switching it to other node immediately but first it opens login screen of other node then it allows to continue from where you left.
I do not want this login screen to come up for other node when anyone of the node is down.
Is there anyone know whether whatever is happening is normal or Is there something i am missing.

Hi,
Whether any one has achieved high availability in MWS.
What i mean is if via load balancer if your request is routed to node1 and while you are working on it node1 goes down and load balancer switches your request to node to then it should be transparent to user what happened at background. It should not ask for login and if on node1 you were searching transaction with filter with certain “DocumentType” then after failover on node2 it should show the search result in transaction for same “DoscumentType”.

I hope someone would have done this already and can guide me here.

your issue looks like related to sticky sessions.
If you have an empower log in this might be helpful
https://empower.softwareag.com/sl24sec/SecuredServices/KCFullTextASP/viewing/view.asp?cstno=8104803&prdfamily=webMethods&KEY=435718-1129003&DSN=PIVOTAL&DST=SR
What is your wm version?

Hi Anjni,
Thanks for your reply. I am using wM8.2.1. It does not seem to me a sticky session issue. As sticky session mean if request comes to node one then it will remain on node1 and that is happening properly. What i want to achieve is complete session replication in case of failover.

what is the fix level?
This was resolved in MWS_8.2_SP1_Fix2.

Hi,
MWS is on FIX 14.
Can you tell me in FIX 2 which issue was related to my issue and got fixed?

If its possible can you share your apache LB conf file?

Hi,
Please find the attachment.
httpd.conf (18.5 KB)

MWS does not support session replication to another node. If the cluster node where your sticky session was originally established goes down then that user will be required to login again since your session information will have been lost.

If you want to make a feature request to support session replication in a future release, you should make your request known @ http://brainstorm.softwareag.com/BrainstormHome

Hi Eric,
Thanks for your reply. At least from your post one thing is clear that MWS does not support what i am looking for. Till now i was thinking that it works in MWS and I am not able to do it.

Hi,
Whether anyone has ever achieved session replication in case of failover by any mean.