NimbostratusBIG-IP DNS Manual Resume
Hi, We host a dual data center solution on behalf of our customers which takes advantage of the BIGIP-DNS and LTM in order to provide internal load-balancing and site to site failover. We use the manual resume feature on the BIGIP-DNS pool to ensure that should a service fail manual intervention is required to enable it.
Over the weekend our hosting provider suffered a power outage, and as a result our service failed over to the remote data center. All was well until power was restored, at which point all the equipment at the failed data center recovered (including the BIGIP-DNS / LTMs), however traffic was directed to the data center unexpectedly. After looking into this further I have discovered that when a node is disabled using this feature, the state change is not replicated to sync group members or stored in the configuration, which explains the behavior, see: -
https://support.f5.com/kb/en-us/solutions/public/k/10/sol10640438.html?sr=57090903
With this in mind, do you have any suggestions as to how to achieve the desired behavior, in that the service remains disabled in these circumstances? I've considered the "time until up" setting on the LTM monitors, will the service appear as down to the BIGIP-DNS System until the "time until up" interval completes, and if so will this cause the BIGIP-GTM to disable the pool member after a restart before directing traffic to it?
