What is DR? How do you define it? I would really like to know. Please put your thoughts in the comments but a recent event has made me re-think what DR truly is.
I have a SQL Server that is very mission critical and if the database is not up and accessible the company is losing money. This is very important and needs to have as close to 100% uptime as possible since this application is part of a consolidation project and that means that the server is taking transactions just about 24 hours a day.
This is a server that I inherited and it seems like the vendor has had control of the environment pretty much exclusively. I have been brought in and I have made some recommendations to help improve performance and I have made some comments about HA and DR. I have made comments because the vendor is essentially using Transactional Replication as a solution for both. They do not have Primary keys in some tables so that means we do not have those tables on the replicated servers. They also are not replicating Indexes, Triggers, and Stored Procs. This would be OK if they had a system in place to script them out and apply them on the replicated database but they do not.
So, Something happened that caused kerberos to break and now we can only use sql server authentication to the instance. We only have one SQL Server account and that account is dbo in the production database and has no other roles except public. The SA account has been disabled, and on top of that no one knows the password for it.
So to me this is a disaster. I also forgot to mention that the Agent will not start up. I do not think that a cluster would have helped us since the server/instance name for the cluster would be the same even when failed over to different hardware. So I classify this as a situation where DR needs to be in place.
What say you about what a Disaster is. I do not think it is when your entire data-center is taken out anymore. It could be just about anything and if you have not planned for it then you will pay the price as I did.