Forum Discussion

Otto_Schmedlap_'s avatar
Otto_Schmedlap_
Icon for Nimbostratus rankNimbostratus
Aug 14, 2006

best practices for load simulation of iRule

I apologize if this is not the correct forum for this posting.

 

 

With the logic complexities of our iRules we need to do load simulations to avoid adverse system level conditions in production.

 

 

I would like feedback for what/how others test their iRules and what to watch in F5 and system statistics, when testing iRules.

 

 

Thank you.

 

  • Colin_Walker_12's avatar
    Colin_Walker_12
    Historic F5 Account
    Well, when testing my iRules I generally pass benign traffic at first. I try to start out with traffic that I know fits the profile of what the rule is looking for so that I can do a baseline kind of test...nice and easy. Once I note that the rule is working as expected, then it's time to start throwing the curveballs.

     

     

    Pass some traffic that has unexpected characters, perhaps RST in the middle of the connection and see how the next connection is handled, that sort of thing. This is more of a logic test than a stress test.

     

     

    If you're looking for straight load simulations...then all you really need to do is get some machines to dump as much traffic as possible (likely working up in incremental jumps) through your BIG-IP on the VIP with the rule applied, and watch the resource usage on the BIG-IP (Memory, CPU, Overall Load, etc.)

     

     

    HTH,

     

    Colin