Summary

Disaster recovery is a very important part of not only your migration plan, but also your day-to-day operations from the standpoint of AD as well as your ProLiant servers. This chapter has described various aspects of disaster recovery from the loss of a single DC to the loss of an entire forest, and recommendations were provided. In addition, important issues were discussed surrounding the restoration of AD objects and the back link issue where users lose membership in cross-domain groups. At this writing, this is a very new understanding of the impact of restoring objects. The techniques for recovery described here are important in order to affect a full recovery.

Disaster recovery encompasses actions related to hardware and software. These actions must be designed and tested in order to have the enterprise prepared for a situation where loss of hardware or corruption of software might occur. Developing and testing such a strategy will minimize downtime and the associated financial and perhaps legal risk posed by loss of data.

I highly recommend that you develop a disaster recovery plan, and thoroughly test it. It doesn't seem important until you need it. I have talked to scores of Administrators who have a great backup procedure, but don't regularly test the restore capability. If you can't restore the data, you are wasting time and money on making backups . Likewise, you should develop and test a recovery of the forest ”not only in case of a natural disaster wiping out all of your DCs, but also in case of schema corruption that could be cured by recovering the forest. Make sure you engage experienced people in developing and testing this plan.