Stickyness on ACE and CSS 11500

Unanswered Question
Dec 11th, 2009


I've configured the following on ACE and CSS11500 respectively (left out the unimportant lines in the content-rule):

sticky http-cookie WPS6 sticky-nbp-wps6

  cookie-insert browser expire

  replicate sticky

  serverfarm sfarm-nbp-wps6 backup sfarm-nbp-wps6-sorry


content sb-wps-inside

    advanced-balance arrowpoint-cookie

    balance weightedrr

    arrowpoint-cookie name WPS6

    weight 9


Are there any major differences in how the ACE and the CSS handles cookie-based stickyness based on the above listed configurations? From what I've read, the CSS looks for the cookie-value at each GET and forwards traffic based on the value in the cookie. Only when the cookie expires, will traffic be subjected to the LB-algorithm again. But how exactly does the ACE handles this? Does is look for the cookie value at each GET and match it to the corresponding entry in the sticky-database? Or is this method only applied, when persistence-rebalance is configured? If I manually delete the cookie in the browser, will that in any way affect how the ACE handles the subsequent GET or will it simply reuse the sticky parameters applied at the first GET, until the cookie expires and is removed from the sticky-database?



I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Gilles Dufour Mon, 12/14/2009 - 06:43

On ACE, without persistence rebalance, we look at the first GET and that treat every subsequent packets as data ...

If you enable persistence rebalance, we look at each request separately.

If the cookie is not present inside one request, we loadbalance the request and send a new cookie.


UHansen1976 Mon, 12/14/2009 - 06:52

Hi Gilles,

Thanks for the reply.

I had a feeling that persistence-rebalance could be the answer. So, would it be safe to assume, that the 'advanced-balance arrowpoint-cookie' feature on CSS is in someway similar to persistence-rebalance, in that it looks for the cookie at each GET and re-loadbalances, if it fails to detect a cookie?



This Discussion