05-24-2016 03:19 AM
As part of a supplier datacentre failover, it was noticed that traffic was still be directed to the original datacentre.
The culprit was/is beleived to be the DNS!Max_TTL, DNS!Min_TTL settings, and the default 24hr value. (the records in question had a dns query response TTL of 300-600 seconds)
However, in evaluating the impact changing these settings have, we noticed that the DNS server configured on the ADC was recieving a DNS request at least once every 60secs, no matter what these values were set at.
Changing these settings never resulted in the queries to the DNS server being greater than at minute intervals. (also some queries were repeated a few seconds apart)
It was proven that changing a DNS record would affect the IP address the ADC uses once the new query had responded with the new address (tested incase there was some higher level ADC DNS cache above the OS).
So, really confused at how these settings are meant to work, whether they do, and if they affect both pool members and traffic-script calls.
NOTE: the above tests were done with pool member DNS names only, on version 10.4 of the VMware VTM vappliance.