We have all of our systems set up to do internal DNS to a local name server and a secondary name server so that in the event the local DNS server is down they should go over the WAN to the secondary that is up in a remote location.
During a recent maintenance window the primary name server locally was taken offline and our UNIX/Linux systems with Veritas Cluster (VCS) and Red hat cluster both experienced issues with cluster resources.
Trying to determine how long resolution should be delayed when the servers have to go to the secondary (i.e. how long is the timeout to the primary) as it seems to hang for a while if the primary server is there but hung up? Is the length of time it waits before moving to the seconday tunable? This is for Solaris 10 and RHEL 6.