Server Load balancing using Cisco Content Switching

I am trying to load balance between 4 web servers using Cisco Content Switch (CSS) hardware and software. I have put the configuration in place for the frontend, and backend side of the CSS, and have configured the circuits, and interface VLAN on the CSS.  
What I would like to know is whether I need to configure "IP Redundant-interface" as VRRP for the servers.
aagoussiAsked:
Who is Participating?
 
cstosgaleConnect With a Mentor Commented:
It is a bit confusing but bear in mind what you are trying to configure is pretty advanced stuff!

If say you have two web servers you want to load balance behind the CSS.

1. Create a service for each server using their real IP
2. create a content rule as the example below:-

owner arrowpoint
   content ruleWap
     vip address 192.168.128.151
     protocol tcp
     port 80
     url "/*"
     add service server1
     add service server2
     header-field-rule wap
     active

This contentrule has a virtual IP. This is the IP that clients access. It then sends the client requests to the real IPs of the server. I believe it does this using NAT, so usually yours servers are in one subnet off an "inside" interface and your virtual IP is on an "outside"

When you have two CSS boxes, you need to have each box share the addresses the CSS has on either side. The ip virtual-router command is for the inside and the ip redundant-vip command is for the outside.

This is because:-

For the servers to get out to the rest of the network, they need to go to their default gateway. In this case, this is the CSS boxes. In order for this side to be redundant, VRRP is used and a "virtual router" is created.

On the other side, the CSS box that is active needs to know that it is responsible for replying to requests for the vip. This is VIP redundancy because if one of the boxes goes down, the other one will take ownership of the vip.

here is a link to a guide explaining vip and virtual interface redundancy: http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11500series/v8.20_v8.10/configuration/redundancy/guide/VIPRedun.html

Its a bit mind bending at first, but as I say this is pretty high level enterprise stuff!

the simple answer to your original question is yes you do, as per the guide above.
0
 
cstosgaleCommented:
Hi,

only if you have more than one content switch. VRRP is for redundancy between content switches. However, I would strongly recommend aving at least two content switches as otherwise you will still have a single point of failure!
0
 
aagoussiAuthor Commented:
In my environment I have two CSS pairing between each other.  I thought that two CSS pairing together for redundancy will use the commands below:
"ip virtual-router" and  "ip redundant-vip" ?

I also thought that servers loadbalancers will use the command " ip redundant-interface" for redundancy between each other?

In my environment I use the three commands above, and I am pretty much confuse about the whole thing.  I tried to ask techies around, and read the Cisco manuals on CSS, but no one seems to have a pretty good idea how to explain it in a clear and concise way.
0
 
aagoussiAuthor Commented:
cstosgale,
You are right by saying that this is advanced stuff.  Unfortunately they aren't many people around who understand the concept.
I have increased the point to 500 due to complexity.
I am now going to digest the document you sent from the cisco web site, and also look at your solution, and see how it goes down. Once the concept is clear I will accept your solution as final which is very helpful.
I will speak to you or other expert sometime tomorrow.
0
All Courses

From novice to tech pro — start learning today.