|
On Thu, 3 Feb 2005, Bilia Gilad wrote:
Hi i configured oracle application server 10g with lvs .
All works fine except oracle portal .
In the oracle portal cluster manual :
http://www.tju.cn/docs/oas90400/core.904/b12115/networking.htm
" The Parallel Page Engine (PPE) in Portal makes loopback connections to Oracle
Application Server Web Cache for requesting page metadata information. In a default
configuration, OracleAS Web Cache and the OracleAS Portal middle-tier are on the same
machine and the loopback is local. When Oracle Application Server is front-ended by an
LBR, all loopback requests from the PPE will start contacting OracleAS Web Cache through
the LBR. Assume that the OracleAS Portal middle-tier and OracleAS Web Cache are on the
same machine, or even on the same subnet. Then, without additional configuration,
loopback requests result in network handshake issues during the socket connection calls.
"
" In order for loopbacks to happen successfully, you must set up a Network Address
Translation (NAT) bounce back rule in the LBR, which essentially configures the LBR as a
proxy for requests coming to it from inside the firewall. This causes the response to be
sent back to the source address on the network, and then forwarded back to the client.
"
Does any one know how can I make the bounce back rule in the LBR( load balancer
) ?
Does this mean We have to work with NAT ( we prefer DR )?
We run LVS-DR with Oracle portal and don't experience any problems. When the
application server connects to the VIP the connection is internal to the
server itself as each server is configured with the VIP on lo:1.
|
