how CSS 11501is load balancer

[sidris-sid]

Hi:

I have installed Cisco load balancer “CSS 11501”,I need to understand what is the method that will be used to distribute the traffic among let’s say 3 different servers “each server is stand alone” how can be done?

Is there a way that I can monitor the traffic behavior so I can tweak it by percentage for example all Servers should have same priority.

In case one of my servers is having http or https issue & I need my CSS appliance to understand automatic to remove this server out of server’s HTTP domain.”I don’t need to respond with http error to users”

Waiting for your reply.

Regards,

Saeed

[kitanaka]

Hello, sidris-sid

By default, CSS does roundrobin accesses from clients to services (real servers) associated the content rule. Pls look at page below.

http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11500series/v8.20/configuration/content_lb/guide/ContRule.html#wp1038118

You can assign different weights on each services so that accesses from clients are load-balanced according to assigned weight.

http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/css11500series/v8.20/configuration/content_lb/guide/Services.html#wp1050713

You can monitor hit count for each services under "show summary" output as below.
I assigned weight 1 for S1, weight 2 for S2 and weight 3 for S3. Then initiated 100 accesses.

CSS11501# show service summary

Service Name                     State     Conn  Weight  Avg   State  Load  Transitions

S1                                         Alive         0           1          2           18
S2                                         Alive         0           2          2           14
S3                                         Alive         0           3          2           10

CSS11501# show summary        
Global Bypass Counters:
   No Rule Bypass Count:     0
   Acl Bypass Count:         0

Owner            Content Rules    State     Services         Service Hits

OWNER            CONTENT1         Active      S1               16
                                                                          S2               34
                                                                          S3               50

You can utilize http keepalive to detect failure of application layer (web server)
and take the failed service out from the available list automatically so that new access requests
will not be load-balanced to the failed server.

http://www.cisco.com/en/US/products/hw/contnetw/ps792/products_configuration_example09186a008009452c.shtml

You can create globl keepalive and attach it to services like the link above also you can configure keepalive
individually for each services as below.

CSS11501#  show run service

!************************** SERVICE **************************
service S1
  protocol tcp
  port 80
  ip address 192.168.11.11
  keepalive type http
  keepalive method get
  keepalive port 80
  keepalive uri "/index.html"
  active

service S2
  protocol tcp
  port 80
  ip address 192.168.11.12
  keepalive type http
  keepalive method get
  keepalive port 80
  keepalive uri "/index.html"
  weight 2
  active

service S3
  protocol tcp
  port 80
  ip address 192.168.11.13
  keepalive type http
  keepalive method get
  keepalive port 80
  keepalive uri "/index.html"
  weight 3
  active

CSS11501# show run owner

!*************************** OWNER ***************************
owner OWNER

  content CONTENT1
    balance weightedrr
    vip address 192.168.10.100
    protocol tcp
    port 80
    url "/*"
    add service S1
    add service S2
    add service S3
    active

CSS11501#

Regards,

Kim