Webaccess agent failover from the secondary node to the primary node requires relogin for the user

Webaccess access agent running on a two node cluster on OES2 Linux
Application and the webserver running on win2000 using IIS
Automatic failover of the webaccess agent not happeining eventhough the groupwise provider object is properly configured in console one
Need to manually edit the webacc.cfg file to have the ip address of the webaccess agent node
Failover from the secondary to the primary node needs relogin for the user

Reported to engineering

Steps to duplicate:

1) Configure webaccess agent for clustering in a two node cluster
2) Confirm that the groupwise provider object is properly modified to have the information of both webaccess agent nodes
3) Modifications made to the groupwise provider object does not get reflected in webacc.cfg file and the failover not happening
4) Modify the webacc.cfg file to have an entry for the ip address of both the webaccess nodes : For that edit  the webacc.cfg file and search for
     "Provider.default=GWAP" section and just above that line add the following lines:
     Provider.GWAP.Default.address.1=IP address of the first webaccess agent node:port number used by the webacess agent node
     Provider.GWAP.Default.address.2=IP address of the second webaccess agent node:port number used by the webacess agent node
     Continue this till you have all the ip addresses listed .
5) Restart webserver , application and the webaccess agent
6) This starts the failover fine from the primary node to the secondary node and the user do not have to relogin to webaccess , but the failover from the secondary node to
     the primary node , needs relogin for the user in the webaccess