05-05-2003 06:19 AM - edited 03-02-2019 07:07 AM
I have a CSS 11000. I have layer 3 load balancing set up and thought it monitored http ports as well as ping availability. When a server goes down it takes the server out of the round robin, or so I thought. I had an incident this weekend, and the server didn't go down but the http application froze but the server didn't take the server out of the round robin. is there additional settings I need to put in place to monitor http availability??
05-05-2003 06:07 PM
If you don't specifically specify a keepalive type, the CSS uses ICMp by default. since your server would have still been pingable, then it wouldn't have know it had gone down.
There are many keepalive types, http being one of them. For a www server, it's probably best to use a http keepalive.
See http://www.cisco.com/en/US/partner/products/hw/contnetw/ps789/products_configuration_example09186a008009438d.shtml for an example.
05-06-2003 12:08 PM
Thanks for the response. I made the following changes for each service:
keepalive type http
but after doing so all services were listed as down when doing a show services. is there something i am missing as the http applications are working if i go directly to each server, just not through the rule, obviously because they were taken out of the round robin because they were listed as down via http keepalives.
thanks
05-06-2003 12:10 PM
the http keepalive method is set to HEAD. whats the difference between HEAD GET and POST.
thanks,
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide