Startup and Shutdown of high Availability Oracle Database Server

In a 24*7 running environment, the terms "startup" and "shutdown" might appear redundant, since a system is supposedly never shut down in such a situation. As any experienced administrator knows, however, these are inevitable events that happen either intentionally or accidentally. Hopefully, in such situations, database machines that support mission significant application will have adequate failover mechanisms in place to replace the primary systems and reduce the dependence on them when they go down. In the absence of a standby system, however, the speed of restarting a database server after it goes down assumes great importance. Even at sites where there are standby systems, albeit less powerful that the primary database machines, prolonged periods of process under the standby database systems may not be acceptable to the organization.

Accordingly, it becomes critical to restart the primary systems as fast as likely. Additionally, if the time taken to bring up the primary database server lower than or equivalent to the time taken to switch over to a standby system, it may not be meaningful to resort to using the standby database server. Only in conditions in which the primary system experiences a serious failure and is expected to remain unavailable for periods longer that the time it takes to switch over to the standby database, is the switch worthwhile. The longer the time is taken to restart a system, means the higher the Mean Time Recovery.

There are other things that would affect your bottom line whether to move around to the life data source due to the main data source taking a while to come up. The main data source would need to be resynchronized with the life data source, after it lastly comes back up. This resynchronization may take a while. Furthermore, the life data source may be managing on a less highly effective program, and once the main data source becomes available, another switchover would need to happen from the additional to the main to avoid lengthy times of managing on the less highly effective program. This switchback would require another data source rebound at the very least, thus including to the final recovery time. Last, the life would have to be completely refurbished to go with main, including a lot of perform. If you are using Oracle dataguard then it should avoid this kind of ridicules perform.