Having a power failure can create many other problems up to and including a server failure. In the event of a power failure, you need to do the following:

1. Ensure the server is cleanly and properly shut down.
2. Wait for at least 15 minutes after the power has come back on.
3. Power on the server and verify that all hardware is intact and no failures have happened
4. Verify the network connectivity
5. Verify that all services started properly
6. Verify the integrity of all applications
7. Verify all databases are intact and none are marked as suspect
8. If everything looks fine, open the server for production

Server failure

There are several options to recover from a failed server if it can be determined that the drive arrays are still intact and no corruption has occurred. The order of preference for these options is:

1. Replace the failed component with a new component
2. Replace the failed component with a component from a non-critical server
3. Swap the arrays from the failed server into the cold standby server
4. Restore an image of the server to the cold standby
5. Restore from tape

Replace the failed component with a new component

1. Diagnose the problem and determine which component has failed.
2. Locate the number for your local hardware vendor.
3. Request the necessary component on an urgent basis.
4. If the component will take longer than 1 hour to arrive go on to option 2.

Replace the failed component with a component from a non-critical server

1. Diagnose the problem and determine which component failed.
2. Locate a non-critical server that contains the same component
3. Shut down the non-critical server
4. Remove the component and install it in the failed server.
5. Power up the server
6. If you have another failure, something additional is wrong and you need to proceed to option 3.

6.5 Disaster Recovery Plan 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18