We have a production SQL Server database server shipping transactional log backups to two standby servers. The disaster recovery plan is already finished: we have a well documented procedure and people trained to put the standby server into production, and initiating replication, enabling the jobs, etc, with minimal downtime.
The problem that is gaining discussion is not the contingency plan itself, but the DECISION to put the standby server into production and losing, in the worst case, 12 minutes of information (the transaction log backup runs every 10 minutes and is very fast to be copied to the other servers).
The decision could be difficult because we can waste time trying to identify the problem. On the other hand, the problem could be simple to resolve and we could put back the server into production without using the other servers.
We understand that the situation will become very stressful in the event of a system failure, and we think that in these situations, it is better to have a standard procedure and a minimum of decisions.
So, we have a dilemma. Is it better to just change servers when something goes wrong with the main server, or better to try to identify and resolve the problem in the main server? What do you guys think about this?