Forum Discussion

bkchng_148728's avatar
bkchng_148728
Icon for Nimbostratus rankNimbostratus
Apr 17, 2014

Understanding the GTM's Return to DNS Load Balancing method

Hi,

 

I've noticed that when Round Robin is configured as the preferred load balancing method at the Wide IP level (I have 2 pools), and Return to DNS is the fallback method (as per default settings), the statistics show quite a number of "Returned to DNS" results (about the same number as the "Preferred" results), even though both pools are up.

 

The document says "The Return to DNS load balancing mode is another special mode that you can use to immediately return connection requests to the Local DNS for resolution. This mode is particularly useful if you want to temporarily remove a pool from service, or if you want to limit a pool in a single pool configuration to only one or two load balancing attempts.". But I'm still wondering why there are attempts to "return to DNS"; perhaps it's because I don't fully understand what this does. Could someone shed light on this?

 

Thank you so much for your help!

 

BIG-IP DNS
deployment

1 Reply

Sort By
  • James_Thomson's avatar
    James_Thomson
    Icon for Employee rankEmployee
    Apr 17, 2014

    If you have a widieip www.site.com - pool A - serverA - pool B - serverB

     

    At the wideip level there is no fallback method. So, you may have Round Robin at the wideip level which means GTM will round robin requests between pool A and pool B. Then, in the pool, there are 3 levels of load balancing: Preferred, Alternate, Fallback Defaults are RR,RR,Return to DNS

     

    In that configuration, Return to DNS should only ever be used when all the pool members are marked down (red diamond) or "unknown" (blue square). In your two pools, what are those 3 methods set to? I've seen a scenario where I'm using two dynamic methods as the first two and Return to DNS as my last and if GTM doesn't have enough dynamic info to make a decision, it drops down. If my pool members are generic hosts and I pick "Least Connections" and "VS Capacity", then Return to DNS. Those first two require intelligent Communication with an LTM instance. So, it doesn't have enough info to make that decision and will drop to return to DNS.

     

    You should look at your log files and see if you have any of those pool members flapping (going from green to red and back).