Load balancer help setup needed

Good Morning, Steve.

What URL are you using to do the XML POST to the LineRate VIP? Generally, you would want to craft your post to the VIP just the same as you would directly to the server. For example, with your above URL example, you might send the XML POST request to the following URL (presuming you configured your VIP to listen on port 7800):

Essentially the LineRate proxy will then choose 1 of your 2 windows servers to relay the request to. One other thing that you may want to be conscious of here is persistence. If your web service requires authentication or saves any kind of state in the application, you likely will want persistence enabled on the Virtual Server in LineRate. What this does is ensures a given client talks to the same backend server so that he doesn't, for example, make a successful request to authenticate to Server 1 and then get his second request sent to Server 2 who doesn't know he exists. You can read more about this here: Virtual Server Persistence

Hope this helps.

-Andrew

Hi,

Thanks for getting back Andrew. Just tried that and I get..

504 Gateway Timeout

Steve.

Right...Think I have solved it..... I noticed that my Virtual IP status was UP but had no associated interface. So I googled that and found.. https://devcentral.f5.com/questions/linerate-version25-virtual-ip-status-up-but-no-associated-system-interface-routed-traffic-only

I needed to create a virtual IP address that was associated with an interface. So I set my Virtual IP address to be the same as Interface em1.

Then I got the response I was expecting.

Now I just need to make sure that it is actually load balancing. And what happens if I switch one off?

Steve.

Problem now is that its not load balancing. If I shut down one of the Integration Servers, I get a 502 bad gateway error every time I POST xml. Or if I shut the other one I still get through every time.

So its only going to one of the real servers??

Its probably because its all on the same subnet.

Any ideas?

Steve.

Thought I would post my interface config..

interface em0
 mtu 1500
 ip address 10.224.100.200 255.255.255.0
!
interface em1
 mtu 1500
 ip address 10.224.100.201 255.255.255.0
!
interface em2
 mtu 1500
 ip address 10.224.100.202 255.255.255.0
!
ip route 0.0.0.0/0 10.224.100.1
!
health-monitor fghiib01uat
 type http
 admin-status online
!
health-monitor fghiib02uat
 type http
 admin-status online
!
ssl profile self-signed
 attach certificate self-signed
 attach key self-signed
!
real-server fghiib01uat
 ip address 10.224.100.103 7800
 base iib01uat_web
 admin-status online
!
real-server fghiib02uat
 ip address 10.224.100.104 7800
 base iib01uat_web
 admin-status online
!
virtual-ip vipiibuat
 ip address 10.224.100.201 7800
 base vipiibuat
 admin-status online
!
real-server group rsgroup_websilo1
 members by regex "fghiib.*"
 member fghiib02uat
 member fghiib01uat
!
virtual-server websilo1
 service http
  persist source-ip
 attach virtual-ip vipiibuat default
 attach real-server fghiib02uat weight 0
 attach real-server fghiib01uat weight 0


Code

Its definitely not load balancing. I have set different messages to respond from each server. If I set the admin status on server 1 to offline, then it gets a response from server 2. But when I bring server 1 back online it continues to only get responses from server 2. Then if I take server 2 offline it gets responses from server 1. Then when I bring server 2 back online it continues to get responses from server 1.

So it only ever sends to one server. But it can obviously send to both, so why is it not load balancing using round robin?

Steve.

That worked perfectly...

Thanks very much Andrew.