IBM Load Balancer and Cluster IP

andridg used Ask the Experts™
We have completed an installation of 3x WAS IBM HTTP servers running on an I570 Server. Separate from this is an IBM X3550 Edge Server, running load balancer program as well. The Edge server has an IP address on the physical NIC and the Load balancer Server has a cluster IP which is a virtual address.  The workstation in our LAN .../23 can all see the ping the .220 & .229 IP.  However the workstations on the WAN can only intermittently see the IP of the cluster, which is .229.  They can see without issues ping the physical nic of the Edge server (.220).  I have check my routing, and the routing of each of the workstation outside the WAN, they all point to the correct Server. I do for example a ping to the same server with both IP's and the result can be seen on the attachments.  Please help because this is coming to a point that it's becoming a big frustration.
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
HonorGodSoftware Engineer

What else is on your network?

What is the (exact) topology?

Do you have any bridges/switches/routers/firewalls or appliances between the machines?

Have you checked the configuration for these network entities?


Hi, We have a core cisco 6500 and all servers are on one LAN, we have a cisco router inside our LAN and then a FW which connect to the Internet.  These addresses that we are using has not been used by any other servers at all, that is why we have chosen them for use, there is not conflict with these Server IP's.  Our topology of the network is BUS
HonorGodSoftware Engineer

Sorry for the confusion.  By topology, I meant a diagram that shows the nodes (physical machines) in your network, and the connections between each.  For example, you may have a virtual network that has a large number of intermediate network segments between your nodes.  Does this make sense?

When you say "We have ... an installation of 3x WAS IBM HTTP servers..." I am not exactly sure what you mean.  Are you saying that you have 30+ IHS (IBM HTTP Servers)?

What version of IHS is being used?
Are they all at the same level?

Does this make sense?
Hi There,
We have 2x class-C addresses /23 in our network in HQ, they are all connected to a single Cisco Router inside the LAN, behind the FW.  These nodes are connected to Cisco switches on each floor and then have fiber running to the Core switch, from the core switch we have our internal router connected.  From the internal Router we have multiple lease lines connected to the branches., see attachment (228574.jpg)
Let first forget about the WAS IBM HTTPS servers, and lets look at the Load Balance Server, which is the X3550 which is mentioned above.  The clients on our WAN's can for example do icmp to that Load balance server, however the results are as shown previously on above postings.  As mentioned prviously, the LB (Load Balancer has a virtual cluster address and a physical IP on the NIC)
This problem only resides on our WAN's, not on the HQ LAN, however we have looked @ our Router,  and all seems like it's OK.
Hope this helps


Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial