Forum Discussion

Melissa_Good_10's avatar
Melissa_Good_10
Icon for Nimbostratus rankNimbostratus
Oct 31, 2013

Fall back host - how to find out what condition is causing it to invoke

I have a pair of 3900s running 11.2 code. I have a web application that takes requests from the internet using Akamai caching - after a recent code change we have seen a high rate of invokes of the 'fallback host' - and we are wondering if there is a way to find out what condition is causing it. We have checked the obvious - the pools were not down and there is nothing in the logs that would indicate that the pool members had come in and out of service. I am not seeing any RST's issued to the load balancer. Anyone have any ideas?

 

application delivery
availability
deployment
devops
iRules

3 Replies

Sort By
  • What_Lies_Bene1's avatar
    What_Lies_Bene1
    Icon for Cirrostratus rankCirrostratus
    Nov 01, 2013

    You've a few options, assuming the fallback is specified in a HTTP profile;

     

    1) Run a packet capture (as a background process) and write it to a file; inspect it closely with Wireshark or a packet analysis tool if you've one available.

     

    2) Create an iRule which logs data under the LB_FAILED event. See here for info on some of the data you can collect and log: https://devcentral.f5.com/wiki/iRules.LB.ashx

     

    BTW, what do you have configured for the Pool's Action on Service Down setting?

     

  • Melissa_Good_10's avatar
    Melissa_Good_10
    Icon for Nimbostratus rankNimbostratus
    Nov 01, 2013

    I have a Fluke NTM - but am waiting for taps to bring in the external data. The logging iRule is interesting - we will suggest that on the troubleshooting call so thank you for that suggestion. The pool's action on service down is reselect, specifying 2 x.