Hoping someone can point me in the right direction for a very annoying persistent issue.
DBX points to an MS SQL server cluster using a DNS host name in the connection string. Normally it works pretty well except for when there is a site change and a bunch of scheduled queries time out. The issue can be replicated in SQL Explorer but the timeouts can be quite inconsistent.
Tests on the search head server show that the host name is being resolved correctly and the server can connect to all resolved IPs. I did note the order of the resolved IPs is somewhat unpredictable and the IP for an offline node may be returned first.
My running theory is DBConnect is attempting to connect to the first resolved IP then failing with timeout before it's able to connect to one of the other IPs. Is this expected behaviour? Is there any way to reduce the connection timeout value?
I have found a number of posts on these boards relating to similar problems but they don't quite apply here.